A Comprehensive Study on Next-Generation Electromagnetics Devices and Techniques for Internet of Everything (IoE)

Abstract

Evolution of mobile broadband is ensured by adopting a unified and more capable radio interface (RI). For ubiquitous connectivity among a wide variety of wireless applications, the RI enables the adoption of an adaptive bandwidth with high spectrum flexibility. To this end, the modern-day communication system needs to cater to extremely high bandwidth, starting from below 1 GHz to 100 GHz, based on different deployments. This instigates the creation of a platform called the Internet of Everything (IoE), which is based on the concept of all-round connectivity involving humans to different objects or things via sensors. In simple words, IoE is the intelligent connection of people, processes, data, and things. To enable seamless connectivity, IoE resorts to low-cost, compact, and flexible broadband antennas, RFID-based sensors, wearable electromagnetic (EM) structures, circuits, wireless body area networks (WBAN), and the integration of these complex elements and systems. IoE needs to ensure broader information dissemination via simultaneous transmission of data to multiple users through separate beams and to that end, it takes advantage of metamaterials. The precise geometry and arrangement of metamaterials enable smart properties capable of manipulating EM waves and essentially enable the metamaterial devices to be controlled independently to achieve desirable EM characteristics, such as the direction of propagation and reflection. This review paper presents a comprehensive study on next-generation EM devices and techniques, such as antennas and circuits for wearable and sub 6 GHz 5G applications, WBAN, wireless power transfer (WPT), the direction of arrival (DoA) of propagating waves, RFID based sensors for biomedical and healthcare applications, new techniques of metamaterials as well as transformation optics (TO) and its applications in designing complex media and arbitrary geometry conformal antennas and optical devices that will enable future IoE applications.

1. Introduction

In this modernized era, embedded devices with different types of machinery, sensors, and other utilities connected to the internet have revolutionized and changed people’s convenience in living standards. The Internet of Things (IoT) utilizes the internet or wireless communication to combine all objects, sensors, or machinery in a physical system. The Internet of Everything (IoE) is a concept that emphasizes the Internet of Things (IoT) by focusing on a more complicated system for the machine-to-machine connection that also includes people-to-people as well as people-to-processes connections. It is a super-ordinate of IoT, a domain where many smart objects connect over public or private networks that use specific protocols to produce real-time data. This IoE is based on the combination of three pillars: intelligent objects, information-centric networks, and real-time insights data from sensors [1]. For information-centering networks of IoE, the bandwidth requirement is a crucial aspect [1].

In recent years, wireless communications and IoE have made significant growth. The next-generation wireless technology is the combination of environmental sensing, data processing, and highly efficient electromagnetic devices [2]. Furthermore, it has been observed that there is a noticeable amount of increasing wireless devices in developed cities [3]. One of the significant challenges in wireless communications is high bandwidth devices transmitting simultaneously with short packets. In contrast, the standard requirement is to design less complex, energy-efficient communication schemes. Another vital challenge appears to develop effective methods for detecting a small number of active users among a significant number of possible user devices with erratic transmission patterns [4]. These new challenges require fundamental fresh thinking technologies, which will also favor the newest applications of IoE.

Future wireless communication technologies are anticipated to create a path for the Internet of Everything (IoE) concept, which envisions a network of interconnected devices without the help of any media. Such significant technologies which focus on this increasing research interest and also are involved and implemented in real life have been reviewed in this article. Here, a descriptive overview has been presented on the direction of arrival, theory of transformation electromagnetic or optic for developing meta-material devices, wireless power transfer, sub-6 GHz for implementing 5G communication, wireless body area network, etc. These technologies can be integrated with IoE to gain more bandwidths and data rates for simultaneous wireless information and power transfer. This article will focus on these prominent wireless connectivity technologies that can be integrated into the big IoE picture.

Meta-materials are manufactured structures made up of sub-wavelength elements with unique and beneficial features not found in nature [5]. Recent studies have demonstrated that meta-materials can improve a variety of antenna attributes, including bandwidth, gain, and size [6]. Meta-material has also shown outstanding performance in sub-wavelength focusing, EM metal, cloaking, 5G communication, Internet of Everything (IoE), internet of things (IoT), etc. [7,8]. Furthermore, in [2], it has been discussed how programmable meta-materials and meta-surfaces can achieve huge advantages for simultaneous wireless information transfer, wireless power transfer, wireless energy harvesting, and 5G communication. Furthermore, meta-materials and meta-surfaces have inspired many new electromagnetic (EM) theory concepts and design techniques for EM devices and systems. In Section 2, a detailed analysis has been presented on one of the EM theories: transformation electromagnetic or transformation optics (TE or TO) for the development of meta-material-inspired electromagnetic devices. Furthermore, a novel approach has been discussed on how to design beam shifters by implementing the same TE/TO technique to get the desired direction.

Wireless Power Transfer (WPT); also known as wireless power transmission, wireless energy transmission, or electromagnetic power transfer), transfers electrical power or energy from a power source to one or more loads by generating a time-oscillating electromagnetic field without any help of a physical link or wire [9]. Recently, this technology has gained enormous attention due to its special applications in our daily life to industrial implementation; such as portable electronic devices, biomedical implants, wireless sensors, radio-frequency identification (RFID), unmanned aerial vehicles, remote charging, and powering of electric vehicles [10]. Section 3 presents an overview of the near-field and far-field WPT techniques. The radio-based WPT has a broad impact on the operation of IoE devices in the home, industries, as well as commercial setup [11]. For the radio-based WPT, radio frequency signals carry the energy simultaneously, which allows charging a number of heterogeneous devices effectively, which will help a lot in the case of the IoE industry [12]. Recently using a relatively small antenna array, this WPT technology has shown incredible performance in systems operating in the microwave band (sub-6 GHz) [11]. In the field of small and multiband antennas for IoT devices, antenna booster technology [13] relies on electrically small elements that are able to excite radiating currents in the ground plane of the IoT device. The proposed matching network [13] makes it easy and fast to operate from single to multiband operation with passive or tunable components. In general, in a wireless body area network, the power source or the batteries for the sensors need to be recharged, where WPT, with the help of DOA, can be implemented to charge the sensors placed in any movable object wirelessly.

The direction of Arrival (DoA) estimation has widely spread applications in wireless communication. To get better and smart performance, this technique can be integrated with smart antennas, 5G, wireless power transfer, radar systems, radio astronomy, wireless sensor network, etc. As with the help of DOA the position can also be determined of the object where we need to transfer the information or power simultaneously. Therefore, a lot of research work focuses on discovering the algorithm to detect the direction of the incident signal. Section 4 discussed different DoA subspace algorithms such as MVDR, conventional MUSIC and ESPRIT (LS-ESPRIT), improved MUSIC, SVD-ESPRIT, and modified ESPRIT (T-ESPRIT and MSVD-ESPRIT). The ESPRIT algorithm surpasses MUSIC by providing a less computation-intensive technique. In contrast, the improved subspace algorithm based on the SVD method and reconstruction of the matrices can successfully estimate the coherent signals DoA. However, it has been supported further by using the Toeplitz matrix reconstruction nature or forward-backward exceptional smoothing (FBSS) to provide enhanced performance and better resolution under low SNR.

The modern healthcare system has reached such a point where medical devices can generate data and send it to be recorded, remotely monitored, and controlled. It is expected that the doctors or physicians and the patient’s relatives will have remote access to these data. Therefore, Wireless Body Area Network (WBAN) can be a part of the IoE concept by connecting with the internet via Wi-Fi, ZigBee, Bluetooth, etc. Wireless body area networks (WBANs) are an emerging wireless technology expected to revolutionize the healthcare system by proceeding with real-time monitoring and data analysis from the sensor data [14]. Section 5 aims to report an overview of the concept of wireless body area network (WBAN) with different network architectures, applications, characteristics, hardware design issues, types of devices, and supporting radio technologies, protocols, standards, and existing challenges. With the widespread use of wireless sensors, energy efficiency has become a concern of today’s research. Wireless sensors usually get power from batteries as a power source. However, battery replacement can be unmanageable in some applications and may require considerable time, affecting the monitored process. However, harvesting energy from natural sources for wireless sensors is also possible. Furthermore, a comprehensive discussion regarding different energy harvesting methods has been discussed with alternative energy sources for wireless body area networks to generate self-sustainable electrical power. WBANs are thought to be impossible to implement without 5G connections [15]. In the future, it is expected that more sophisticated applications will be performed with the help of high data rates under the sixth generation (6G) of wireless networks [16].

IoE/IoT has a significant impact on biomedical and healthcare applications such as monitoring patients remotely, hospital and medical waste management, body scanning to detect cancer, smartwatches to analyze depression, remotely nurse assistance, etc. [17]. Integrated wireless technologies with IoE pave the way to make this application a reality. To facilitate the IoE network, one of the rising pervasive technologies is a radio-frequency identification (RFID). It is a short-range wireless technology used in a wireless sensor network that enables the internet to reach out to the physical objects of real life [18]. Furthermore, RFID interconnects with wireless power transfer and body area networks. Many research works have been performed to deliver power wirelessly to RFID devices [19]. Furthermore, the wireless body area network system can be integrated with the RFID technology to achieve higher efficiency [20]. Section 6 presents a brief discussion on how RFID has played an influential role in modern healthcare applications. The block architecture of RFID systems and the characteristic of active and ultrahigh-frequency (UHF) passive RFID tags have been discussed here. Moreover, a number of research works with the application in biomedical science that happened in the past decades have been reviewed thoroughly.

The development of millimeter-wave 5G wireless technology-enabled IoT/IoE devices will revolutionize the electronics and telecommunications fields [21]. This innovation is one of the most significant achievements in wireless communication and IoE. The combination of 5G and IoE will allow higher data rates, more customer capacity, and more coverage with an increasing number of connections. It will be the path of reality for smart devices [22]. Intelligent devices based on the IoE are extending their support for internet connectivity. These devices have been combined with cutting-edge technology to manage and communicate efficiently utilizing 5G wireless technology. The 5G wireless communication provides the most outstanding advantage over previous 4G or 3G technology by providing real-time communication of IoT devices as 5G offer greater bandwidth in millimeter bands [23]. With the help of this 5G IoE, many researchers are currently focused on constant wireless powered communication (WPC) [24], which can be an excellent initiative for the green batteryless network for low profile wireless sensor networks [25]. For the implementation of 5G, two frequency ranges need to be focused on: sub-6 GHz (less than 6 GHz) and mmWave (above 24 GHz). Section 7 focuses on the 5G implementation (sub-6 GHz) where single and multiple antenna element systems have been briefly reviewed.

2. Transformation Electromagnetic/Optics

The recently introduced transformation electromagnetics/optics (TE/TO) technique [26,27] has revolutionized the developments of meta-materials and devices based on these meta-materials. The TE/TO technique provides extraordinary flexibility of designing electromagnetic devices using an appropriate coordinate transformation method. This also provides an exceptional avenue for registering unique and novel wave-material interactions. The TE/TO design approach is based on the key assumption of the form-invariance of Maxwell’s equations under coordinate transformations [28,29]. Another observation is the interpretation of the material parameters ($\epsilon$, $\mu$) in the transformed coordinate system as a set of material parameters in the original coordinate system [30]. Now, consider the time-domain Maxwell’s curl equations:

$$\nabla \times E=-j\omega \mu H,$$

(1)

$$\nabla \times H=-j\omega \epsilon E,$$

(2)

where E and H are associated electric fields and magnetic fields in a simple coordinate system (Cartesian, cylindrical, or spherical coordinates). The divergence equations in the time-domain are:

$$B=\mu H,$$

(3)

$$D=\epsilon E,$$

(4)

where B and D are electric and magnetic flux densities, respectively, and $\mu$ and $\epsilon$ are permeability and permittivity, respectively. Next let assume a Cartesian coordinate system G$\left(x,y,z\right)$ to describe the original space and G${}^{\prime }$$\left(x^{\prime },y^{\prime },z^{\prime }\right)$ describes the transformed space, as shown in Figure 1. The transformation from G to G${}^{\prime }$ can be described by following:

$$\begin{array}{c}\hfill \begin{array}{c}\hfill x^{\prime }=x^{\prime }(x,y,z)\\ \hfill y^{\prime }=y^{\prime }\left(x,y,z\right)\\ \hfill z^{\prime }=z^{\prime }\left(x,y,z\right)\end{array}\end{array}$$

(5)

Under this new transformed coordinate system, the Maxwell’s equations will remain form-invariant as the following [28,29]:

$$\nabla \times E^{\prime }=-j\omega {\mu }^{\prime }{H}^{\prime },$$

(6)

$$\nabla \times H^{\prime }=-j\omega {\epsilon }^{\prime }{E}^{\prime },$$

(7)

The material parameters in the transformed coordinate system are given by following [30]:

$${\epsilon }^{\prime }=\frac{JJT}{detJ}\epsilon$$

(8)

$${\mu }^{\prime }=\frac{JJT}{detJ}\mu$$

(9)

where J is the Jacobian matrix of the transformation from the G$(x,y,z)$ coordinate system to the new coordinate system, G${}^{\prime }$$\left(x^{\prime },y^{\prime },z^{\prime }\right)$ and $J^T$ is the transpose of matrix, J. J is defined as [31]:

$$J=⌈\begin{array}{ccc}\frac{\partial x^{\prime }}{\partial x}& \frac{\partial x^{\prime }}{\partial y}& \frac{\partial x^{\prime }}{\partial z}\\ \frac{\partial y^{\prime }}{\partial x}& \frac{\partial y^{\prime }}{\partial y}& \frac{\partial y^{\prime }}{\partial z}\\ \frac{\partial z^{\prime }}{\partial x}& \frac{\partial z^{\prime }}{\partial y}& \frac{\partial z^{\prime }}{\partial z}\end{array}⌉$$

(10)

A typical TE/TO technique can be summarized as follows [31] and is illustrated in Figure 1:

I.

Determine a known wave-material relation in the original coordinate system, i.e., a plane wave or a propagating Gaussian beam.

II.

Find the volume of space in the original coordinate system and the associated volume of space in the transformed coordinate system.

III.

Define the coordinate transformation you choose to map your original space to that new transformed space.

IV.

Compute the material parameters in the new transformed space using Equations (8) and (9).

V.

Translate the computed material parameters from the transformed space and acquire the desired material in the original coordinate system.

2.1. Unique Electromagnetics Devices Designed Using TE/TO

Beam-Shifters, Benders, and Rotators

The beam-shifter [33,34] is probably one of the finest and simplest examples of how the TE/TO technique can be employed to freely control the behavior of an electromagnetic beam to the desired direction. In this device, the beam is translated in the desired direction as it propagates through a medium. The first step is to find a set of coordinates in which the beam appears to shift its direction in the desired way. Such a coordinate system is shown in Figure 2. A beam shifter with thickness t was considered [34] starting from $x_0$, a coordinate transformation for shifting the beam is as the following:

$$\left\{\begin{array}{c}{x}^{\prime }\left(x,y,z\right)=x\\ y^{\prime }\left(x,y,z\right)=f(x,y)\\ z^{\prime }(x,y,z)=z\end{array}\right.x_0<x<x_0+t.$$

(11)

As a two-dimensional (2D) condition was considered here, the new component $y^{\prime }\left(x,y,z\right)$ is independent of z. There were no materials introduced in the regions (regions 1, 3, and 5 in Figure 2b) where there was no beam-shifting required. In those regions, the Jacobian will be just an identity matrix. The materials were introduced into the regions of up-shifting (region 2) and down-shifting (region 4) as shown in Figure 2b.

Later, this concept was adopted to steer the radiation pattern of a dipole antenna around an irremovable perfect electric conductor (PEC) object before propagating on its original route [35].

In [33], it was shown that a beam could be shifted in a Cartesian coordinate using an appropriate transformation that translates one coordinate in the Cartesian system. Applying the same type of coordinate translation to the azimuthal coordinate in a circular cylindrical system, it is possible to design a device that can change the direction of beam propagation. In [31], the authors employed a Cartesian-to-cylindrical coordinate transformation to design a two-dimensional bender that rotates the direction of beam propagation by 90°.

In [36], it is shown the techniques of TE/TO can be utilized to manipulate EM fields and rotate them in a specified direction. This approach was later used to control radiation characteristics of an antenna element (as shown in Figure 3) or an antenna array in free space and to rotate it in the desired direction, hence realizing a beam-steerer using TO-based media [37]. This specific TO approach results in material properties which require active tuning to achieve beam-steering.

2.2. Electromagnetic Cloaking and It’s Applications in Antennas

The electromagnetic cloak has drawn a lot of attention from the scientific community because it has been shown that it is possible to make someone or some object invisible. The concept was first introduced by Pendry et al. [26], and later was demonstrated by Schurig et al. [38]. In [39], the authors introduced a nice platform to perform the full-wave simulation of electromagnetic cloaking. The TE/TO technique has been employed to transform a space in the electromagnetic cloak. The idea of cloaking is that a given volume of space is hidden in such a way that the observer from the outside will not be able to see the object hidden in that concealed space. This is achieved by surrounding the object that will be cloaked by a metamaterial shell designed using the TE/TO technique, which guides the electromagnetic wave around the concealed object in the metamaterial shell and goes back to its original orientation. In the metamaterial shell, the wave travels more than the speed of light to catch its original trajectory. Thus, for an external observer, it looks like the wave remains unperturbed through the hidden object and the object simply does not exist to the observer.

The concept behind the cloak is to find a coordinate transformation that will take all space smaller than a given radius and shrink it to a point. To facilitate this kind of transformation, it is important to find an intermediate transformation between Cartesian and cloaking coordinates. Furthermore, this process allows a coordinate transformation between two non-Cartesian coordinate systems which was very helpful and instrumental in taking forward other TE/TO works involving source transformations [40,41]. One of the natural choices for this intermediate coordinate system is cylindrical coordinates if one chooses to cloak a cylindrical object. Instead of a cylindrical object, one also can choose an elliptical cylindrical [42] and two-dimensional eccentric elliptical objects [43] to conceal.

Figure 4 shows the full-wave simulations of two-dimensional (2D) electromagnetic cloaking using COMSOL Multiphysics.

Multiple-antenna technologies have gained much attention in the recent past because of the huge gain they can introduce and the channel capacity levels in modern-day communication systems [44]. However, it poses some real challenges, such as near-field mutual coupling effects that can drastically degrade the electrical parameters of the individual radiating antenna in multiple-antenna environments [45]. Various antenna design technologies [46,47] have been proposed to reduce the severe adverse effects of scattering in the multiple-antenna environment. Recent developments in transformation optics (TO)-based electromagnetic cloaking yield a new direction toward solving antenna design problems in a multiple-antenna environment. In [48], researchers demonstrated a shielding technology using an eccentric elliptical cloak to restore the antenna parameters in a multi-element antenna environment.

Electromagnetic Source Transformation

The same coordinate transformation technique can also be applied to a region with sources (e.g., current and charge distributions), where the sources will be transformed along with the material and behave the same way as the original untransformed source [40,41,49]. The use of source transformations yields fascinating opportunities for the design of complex radiation structures by enabling the fabrication of unique structures with engineered material properties, allowing the transformed geometries to mimic the performance of original ones, which is especially useful when physical constraints require spatial changes to be made to the source. One of the first applications of this technique was suggested by Luo et al. [49], who transformed a dipole current source into a completely new distribution while preserving its properties as a dipole antenna. Kundtz et al. [40] took this concept further and introduced an optical source transformation using a “pinwheel” transformation, where they approximated the sheet current of a simple dipole as a volume current and transformed it to a new current distribution using a “pinwheel” coordinate transformation. Based on these pioneering works, several conformal arrays have been proposed [50,51,52]. Popa et al. [50] proposed a conformal array design, where a nonuniform circular array radiates as a uniformly spaced linear array. Kwon [51] proposed a TO-based circular array design, which performs as a series of line sources embedded in a rectangular metamaterial media. However, these designs used “point” charges or electric line sources as the radiating sources in their numerical solution to validate their proposed concept. A more practical antenna source was not considered in these TO-based phased array designs. In [53], for the first time the source transformation approach was used to design a new linear array, where each individual antenna element was transformed from a single dipole element in free space (as shown in Figure 5).

3. Wireless Power Transfer

There are several methods for transferring energy, but commonly for any WPT system, there will be one transmitter device (TX) connected to the power source and a receiver device (RX) to power the load. Based on the distance between the TX and RX, WPT can be classified in two ways: Far-field (radiative) WPT system and Near-field (non-radiative) WPT system [54]. In a WPT system, frequency, distance, transfer power, and efficiency are the key parameters to determine the system’s performance.

In 1899, the first successful practical experiment was done in the history of the WPT system by Nikola Tesla, where he transmitted 100 MV over a distance of 40 m [55,56]. Nowadays, wireless charging can be seen in a lot of devices, hence it was first established by W. Brown in 1964 when he used rectenna to power a model helicopter [57]. Later on, during the 1970s and 1980s, NASA’s project on solar-powered satellites (SPS) achieved significant developments in microwave power transfer [58]. The potential of WPT has not yet been boosted enough, hence there are numerous research has been conducted on the fundamental features to improve the WPT system namely: high-density power devices, efficient rectennas, innovative circuit architectures [59], different resonator circuits, etc. This section will compare different methodologies, applications, and achieved efficiency over distances for near and far-field WPT systems.

3.1. Near Field (Non-Radiative) WPT System

The Near Field WPT system is mostly preferred for short and mid-range power transmission. The near field region can be defined as $\lambda$/2$\pi$, where $\lambda$ is the wavelength of electrical signal [60]. Here, the transmitted energy wavelength is greater than the receiver distance [61]. In the near field region, the electric and magnetic fields are separate, and both fields can exist independently [60]. The power can be transferred by capacitive coupling (via electric field) and inductive coupling (via magnetic field) [54]. Recently, magnetic resonant coupling and strongly coupled magnetic resonance methods have gained higher efficiency for mid-range application [62]. As near-field techniques are suitable for short to mid-distance, they can transfer power with high efficiency. Furthermore, it is non-radiative, which makes it safer, especially for biomedical implants and health monitoring applications.

3.1.1. Inductive Coupling

Inductive coupling (IC) methodology is one of the most preeminent ways when it comes to wireless charged low-powered devices, medical implants, and powering RFID tags [19,63]. It also has the simplest working principle which transfers energy between two coils located at TX and RX end as shown in Figure 6 [64]. When an alternating current passes through the transmitter coil, a time-varying magnetic field crosses the receiver coil, and an electromotive force is induced at the receiver end, and thus the power is transferred to the load [65]. In the case of inductive coupling, the operating frequency is in the kilohertz range, and the transfer distance is shorter than the diameter of the coils (typically from millimeters to a few centimeters (20 cm)) [19].

In 2011, ref. [66] presented three different inductive links, starting with conventional 2- coil to 3-, 4-coil inductive links and demonstrated a circuit-based theoretical structure to compare 2-, 3-, and 4-coil inductive links. They have proposed three coil inductive links based on achieving high power transfer efficiency (PTE) and the amount of power delivered to the load (PDL). At 12 cm distance, the three coils inductive link has shown 37% efficiency at 13.56 MHz and also offers a significant PDL than two or four coil links.

In a similar year, W. X. Zhong et al. [67] presented a WPT system for wireless charging by introducing multiple coils at the transmitter side whereas one single-coil at the receiver side, while the power transfer distance is 1.5 mm. Based on their configuration, multiple devices can be charged simultaneously whatever their position will be on the charging plate. The researchers have achieved 86% to 89% efficiency depending on the worst and best positioning on the charging plate, while the transfer distance is 1.5 mm.

For inductive coupling, the inductance must have high inductance and quality factors. To improve the inductance the authors in [68] proposed a multispiral printed circuit board (PCB) inductor, which stacked geometry will help to enhance the inductance density. To increase the efficiency of short-range power transferring another new technique has been adopted [69] using a multiple-input single-output (MISO) coil system in 2014, specifically for the industrial application of the wireless sensor. The authors tried to mitigate the magnetic flux leakage in the scenario of increasing transfer distance between two coils and compared the result for both single-input single-output (SISO) and MISO coil systems. Their proposed MISO coil showed 1.5 times more power transfer than the SISO system.

The authors in [70] designed a new approach of using a relay coil in a three-coil system, which shows better efficiency with a relatively greater distance than the traditional two-coil method in 2016. They designed a new U-coil system consisting of three coils- a primary, a secondary, and a relay coil. The relay coil is placed perpendicularly between the primary and secondary coils, which passes the energy between them. Compared with the traditional method, the power transfer efficiency of the U-coil improved to 66% over 100cm distance at 85 kHz.

In 2016, Mehdi Kiani and Ahmed Ibrahim suggested a new figure-of-merit (FoM) that combines both PTE and PDL under specific absorption rate (SAR) constraints [71] for biomedical implants. Another approach of using a long slim dipole coil has been proposed in [72] for home applications. Here, the authors have demonstrated transmitter and receiver reflector characteristics properly, and also the optimum secondary term has been investigated properly for high efficiency.

In 2019, another approach [73] has been proposed the implementation of a coil with alternative clock and counterclockwise winding to establish inductive power transfer (IPT) systems at 13.56 MHz. The authors have successfully driven an LED monitor which establishes their proposed IPT system. An elaborate explanation of equivalent circuits, transfer efficiency, and the measurement over different transfer distances has been presented here. Another approach has been presented in [74] where the authors have implemented a dipole-coil-based magnetic coupler with a new circumferential coupling manner. The evolution of the magnetic coupler’s structure has been analyzed here, and also the coupler’s dimensions have been designed in such a way as to meet the misalignment requirement of the system. The system has achieved 89.7% DC-DC efficiency by transferring 630W output power for underwater applications.

3.1.2. Magnetic Resonant Coupling

Magnetic Resonant Coupling (MRC) is the most popular technology due to its higher efficiency for mid-range WPT applications. In this MRC methodology, the power is transferred from a resonant transmitter to a similarly tuned resonant receiver as illustrated in Figure 7 [64]. The most common technique to increase the efficiency of an IC system is to connect a parallel capacitor to the inductor to form resonance at a specific carrier frequency [75,76]. This resonance is responsible for emphasizing the quality factor (Q-factor) which is directly proportional to the power and eventually increases the power transmission efficiency [64]. The goal of this technology is to optimize power transfer efficiency by focusing power transmission at a specific resonant frequency [10,75]. In general the operating frequency for the MRC is in the megahertz range over the transfer distance of 0.5 to 5 m.

To overcome the crisis of the short battery life and also misaligned coupling coils of unmanned aerial vehicles (UAV), the authors in [77] proposed a WPT system that consists of a new type of magnetic resonance coupling coil-array structure. This novel structure can deliver 65.77W output power at a transfer distance of 30 mm. Wang et al. designed a conformal split-ring loop [78] which has a self-resonant frequency. In general, a standard coil has been constructed with a parallel LC circuit where due to this parallel configuration this kind of structure does not show great power transfer performance. The proposed circuit has a series inductor-capacitor connection, and it has remarkably reached 87.9% transfer efficiency over 22 mm.

For dynamic charging three cost-effective couplers (circular, rectangular, and hexagonal) have been designed in [79]; while doing the optimization of all three couplers, the transfer air gap and the operating frequency remains the same, 200 mm and 35 kHz. To transfer the power effectively the load resistance and mutual inductance play a vital role as the efficiency is inversely proportional to load resistance and proportional to the mutual inductance. For the improvisation of the magnetic field, dual open-loop spiral resonators (OLSRs) have been suggested in [80]. Both single and dual open loop-spiral resonators have been investigated here. Instead of one OLSR, dual OLSR has been proposed here to build up resonators’ surface current, which will lead to strengthening the electromagnetic field. The OLSR has been constructed using metal–insulator–metal (MIM) capacitor and spiral loop inductor. The proposed WPT systems showed 70.8% efficiency over a transmission distance of 31mm while operating at 438.5 MHz.

An innovative resonant inductive link has been demonstrated in [81] based on the appearance of magnetic coupling between two resonators. The transmitter or the primary resonator consists of two spiral printed coils placed on top and bottom of the substrate and also connected to the power source. To maintain a better magnetic coupling, the secondary resonators have been constructed with a square split ring, eventually connected with the implantable medical device. A surface-mounted device capacitor is also used in a shunt configuration to get the optimum resonant frequency. The authors in [82] attended the issue of cross-coupling in a multi-receiver resonant inductive coupling WPT system as the output power is significantly hampered due to this cross-coupling. The author designed such a model where the power or coil position information will not be required. To be orthogonal to the phase of the transmitter current, it is important to control the phase of the receiver current, which can be accomplished by inserting the right reactance into each receiver. Here the authors have proposed a switching circuit connected to the receiver. This compensation method receives the information on the phase of the transmitter and can automatically adjust the phase of the receiver current dynamically to the desired phase.

3.1.3. Strongly Coupled Magnetic Resonance

In 2007, a group of researchers from the Massachusetts Institute of Technology first proposed strongly coupled magnetic resonance (SCMR) to light a bulb in a 2m distance with 40% power transfer efficiency [83]. The working principle of this SCMR describes the placement of two high intermediate resonant coils with high-quality factor (Q) between the transmitter and receiver as depicted in Figure 8 [64]. Like the MRC system, a parallel capacitor has been connected with the coil to create the resonance of the coils [64]. The authors in [83] claim power transfer over distances up to 4 times the diameter of the coils which makes it most efficient in all the coupling techniques.

The efficiency of a strongly magnetic coupling system significantly responds to the alignment between the transmitter and receiver. To reduce the misalignment sensitivity of SCMR another approach has been discussed in [84], where two orthogonal coils have been used instead of standard planar coils. The proposed 3-D SCMR system has achieved 40% power transfer efficiency where the angular misalignment range was 360°. For the same purpose, another research work [85] proposed a novel cylindrical SCMR-based principle that has no nulls throughout the 360° misalignment rotation and can maintain a static high efficiency than any other standard SCMR topology. This new system achieved over 40% efficiency and also can be easily implemented in body wearable applications. A conformal strongly coupled magnetic resonance system (CSCMR) has been proposed in [86] by using a 60 mm radius U-loop as an intermediate resonator. Despite the angular position of the receiver all around the U-loop, this CSCMR system can maintain efficiencies of above 60%. It provides a high transfer efficiency of 73% at a transfer distance of 120 mm in the near field.

To enhance the Q factor and power transfer efficiency a multilayer resonator is demonstrated in [87] where extra 2- and 3- layers of printed spiral coils are inserted in the transmitter-receiver of a 4-coil planar WPT system. For further improvement of PTE, a conductive shorting wall has been placed to connect the resonators. The authors investigated that 77.27% to 84.38% of transmission efficiency can be achieved by implementing this proposed design.

Though strongly coupled magnetic resonance WPT systems get a high-quality factor (Q), the system bandwidth is still limited, especially affecting human wearable devices. In this regard, Zhou et al. in [88] proposed a wideband four-coil SCMR WPT system where the loop coils (transmitter and receiver coils) are placed fixedly at the center of their corresponding resonator. In general, increasing bandwidth in a conventional 4-coil system will cause a declining Q factor, leading to less efficient power transfer. However, this proposed model has increased the bandwidth by over 100% without affecting the overall power transfer efficiency. Another multi-band conformal SCMR has been studied in [89], where the different sizes of multiple pairs of loops resonator have been used.

For dual-band, WPT [90] presented an overlapped defected ground structure (DGS) system using two back-to-back coupled single loop DGS resonators. The proposed circuit shows more than 71 to 73% efficiency and the size has been reduced to 50% without changing any other parameters. In 2019, the authors in [91] presented triple-band WPT near field system by implementing defected ground structure (DGS) topology. In this system, two triple-band DGS bandpass filters have been placed in a row, and this structure can create an independent effective inductance at higher frequencies. This new DGS-WPT system can transfer power over 30 mm and has achieved 68% efficiency at 100.8 MHz.

3.1.4. Capacitive Coupling

Capacitive coupling (CC) or electric coupling is one of the earliest approaches to transfer electrical energy wirelessly and was executed by Nicola Tesla in 1891 [92]. The capacitive coupling can be described as a circuit consisting of two capacitors placed in a transmission distance, where each capacitor will have two parallel metal plates [93]. The bottom plates of both the capacitor will be connected to Tx, and the top metal plates will be connected to the Rx, and between the Tx and Rx, there will be a dielectric medium. When an AC (alternating current) voltage is applied to the transmitter a time-varying electric field will be induced across both plates, and thus power can be transferred in the form of displacement current. Though the working principle is simple, it is not the best approach to transferring power. It can transfer power with good efficiency but in a very short range (a few centimeters). Compared to the IC, the CC WPT system does not make excess eddy-current losses, and also there is no concern about the temperature rising in metal [94]. However, the major drawback is that it requires a higher voltage to improve efficiency. Furthermore, covering longer distances requires a larger place to establish the plates, and also there is a chance of misalignment [10].

In 2018, the authors in [95] proposed a capacitive wireless power transfer system for underwater robotics technology. This research work focused on the behavior of coupling co-efficient in the coupler and also on the product of coupling coefficient(K) and quality factor(Q). The authors showed that the parameters that give maximum KQ will show maximum efficiency and so it can be stated that for achieving high efficiency it is very important to handle the coupling coefficient k and the Q-factor effectively. In [96], a group of researchers investigated both IC and CC for stent-based biomedical implants in 2018. According to the authors, though IC and CC are suitable for transferring the power to stent-based implants, IC is safer according to the safety threshold whereas the implementation of CC can deliver more flexibility in case of use.

In the case of charging electric vehicles via a CC system, multiple parasitic capacitances can appear between the frame of the vehicle and the road which will overload the coupling capacitors, and hence lead to poor WPT efficiency. To overcome the effect of parasitic capacitances, the authors in [97] have proposed to implement split-inductor matching networks to simplify the complex parasitic capacitances into an equivalent four-capacitance model.

They have proposed two prototypes for the capacitive power transfer where the output power is 590 W and 1217 W at 6.78 MHz over the transfer distance of 12 cm. For effectively charging the electrical vehicles and also to mitigate the parasitics present, the authors in [98] innovatively designed matching networks by introducing interleaved-foil (TF) inductors. Two types of TF inductors, semi-toroidal interleaved-foil (STIF) inductors and toroidal interleaved-foil (TIF) inductors have been distinguished from evaluating the performance of the capacitive WPT. This paper shows that compared to traditional solenoid inductors, this TF inductor has shown better tradeoffs, quality factors, and self-resonant frequency.

Different single and double LCLC and LC compensated circuits have been studied over the year as shown in

Table 1. Comparison between different near-field WPT systems.

Methodology		Frequency (Hz)	Transfer Distance	Power (W)	Efficiency (%)	Application
IC	2-coil inductive links [66]	13.56 M	12 cm	49.5 m	15	implantable microelectronic devices
	3-coil inductive links [66]			43.4 m	37
	4-coil inductive links [66]			3.9 m	35
	Multiple Input Single Output (MISO) coil System [69]	38 k	50 mm	42 m	30	wireless sensors
	Dipole-Coil [107]	20 k	7 m	10.3		wireless sensors
	Dipole Coil [108]	20 k	3 m	1403	29	powering sensors
			4 m	471	16
			5 m	209	8
	[109]	215.5 k	66 mm	200		underwater applications
	1 MW resonant inverter including 128 m transmitter [110]	60 kHz	5-cm	818 k	82.7	High Speed Train
	Figure of merit [71]	20 M	10 mm	2.2 m		Millimeter-Sized Biomedical Implants
	Optimal Shaped Dipole-Coil [72]	200 k	1 m	150	83.1	Home Applications
	Circumferential Coupled Dipole-Coil [74]	50 k		630	89.7	Charging Autonomous Underwater Vehicles
	U coil [70]	85 k	100 cm		66	low-power applications
	[111]	60 k	7 cm	180 k	85	online electric train
	[112]	465 k	21 mm	1.0 k	92.41	Lightweight Autonomous Underwater Vehicles
	Inductive Link Design [113]	13.56 M		9.2 m	75	Biomedical plants
MRC	array coil [77]		30 mm	65.77	63.44	unmanned aerial vehicle (UAV)
	array coil [114]	1.4 M	10 cm		81	Electrical Vehicle
			35 cm		60
	Two Concentric Open-Loop Spiral Resonator [80]	438.5 M	31 mm		70.8
	Square Split Ring at receiver end [81]	403 M		1 m	5.24	powering the pacemakers remotely
	Two Coils [115]	20.15 k	15.6 cm	1000	96	Electrical Vehicle
	[116]	23 k		20.92		Electrical Vehicle
	dual-receiver textile coils [117]	6.78 M	0.5 cm to 2 cm		91	Body Wearable Applications
	Split-ring Loop [78]	433 M	22 mm		87.9	Radio Frequency Identification
	J-inverters [118]	50 M	38 mm		75
	SIMO coils [119]	20 to 25 M	4.27 cm	0.84	24
				0.98	29
	LCC compensation circuit [120]	85 k	5 mm to 25 mm	1.78 k	86.1	Automatic Guided Vehicles
	Rectangular coils [79]	35 k	20 cm	8000		For dynamic charging of Electrical Vehicles
	hexagonal coils [79]			4000
	Circular coils [79]			6000
	Dual TX and RX [121]	40 k	7 cm	2100	93.6	Electrical Vehicle
SMRC	2 and 3 Layer of printed spiral resonators with shorted wall [87]	13.56 M	10 cm		77.27 to 84.38	small electronic devices
	tuned 4-coil SCMR system [122]	5.8 G	1 mm		1	Miniature Implanted Devices
	repeater loop or U-loop [86]	40 M	120 mm		73	charging electronic devices
	two orthogonal coils [84]	85.7 M	120 mm		40	The efficiency reported when the angular misalignment is 360°
	DGS [91]	100.8 M	30 mm		68	Triple Band WPT
		140.7 M			60
		182.2 M			65
	Overlapped Single Loop DGS [90]	0.45 G	12.5 mm		71	biomedical applications
		0.95 G			73
CC	Conformal Bumper [123]	530 k	60 cm	S1k	90	Electrical vehicle
	Class-E resonant inverter [124]	1 M	0.25–2 mm	9.63	96.3 to 91
	Interleaved-Foil Coupled Inductors [98]	13.56 M	12 cm	3.7 k	93	Electrical vehicle
	Double-Sided LCLC [99]	1 M	150 mm	2.4 k	90	Electrical vehicle
	double-sided, LC-compensated CC WPT [100]	1.5 M	180 mm	100	66.67
	LC-Compensated Topology [101]	1 M	150 mm	2.84 k	94.5	electric vehicle charging application
	[95]	107.7	20 mm		91.3	underwater robotics technology
	[96]	4 M	30 mm		1	Stent-based Biomedical Implants
	multi-modular CC [125]	6.78 M	12 cm	1.2 k	89.8	Electric vehicle charging
	Introducing glass as dielectric medium [94]			1.6 k	96	Electrical vehicle
	split-inductor matching networks [97]	6.78 M	12 cm	590	88.4	mitigate parasitic capacitance while charging Electric vehicle
	square and circular coupling plate enveloped in PTFE [102]	6.78 M	12 cm	146	84	Mitigate the problem of arcing while charging EV
				590	88.4
				1125	85
				1217	74.7
IC + CC	Hybrid WPT [103]	1 M	20 mm	653	87.7	Railway Application
	Space-Saving Coupler Structure [106]	difer for various transfer distances	150 mm		86
	SS compensation topology [104]			1.1 k	91.9	high power applications
	LCL compensation circuit [105]	1 M	18 mm	100	73.6
	LC-Compensated Topology [101]	1 M	150 mm	2.84 k	94.5	Electric vehicle charging

[99,100,101]. In 2018, Ref. [102] this research work focuses on another important issue while charging electric vehicles, arcing. The authors have proposed four designs (square, circular) of coupling plates enveloped in a thin layer of polytetrafluoroethylene (PTFE). Two of their proposed models show kilowatt-scale power transfer at 6.78 MHz.

In recent years, hybrid wireless power transfer gains more popularity for its better performance. In a resonant structure, the IC system needs a capacitor to compensate while the CC system also needs the inductor to tune the system. As a result, magnetic and electric fields will be generated concurrently across the resonant inductor when resonance occurs. Therefore, researchers have combined both inductive and capacitive coupling to form a hybrid wireless power transfer (HWPT) system [101,103,104,105,106] to achieve a stronger coupling ability.

In 2020, the authors proposed a hybrid WPT system [103] combining both IC and CC systems along with a brief concentration on the misalignment condition of the couplers. This HWPT system has achieved 87.7% efficiency by delivering 653 W output power. If the couplers are placed in misalignment in 0 to 270 mm, the maximum output power variation is 8.3%. In the following year, a group of researchers has presented a new HWPT system [106] by introducing a space-saving coupler structure. This novel design has established an effective capacitive coupling loop considering one pair of frame-shaped metal frames instead of two coupling capacitor plates. This HWPT shows 51.1% efficiency at 632 kHz whereas a pure IC system has shown 37.6% of efficient power transfer over a 30 cm distance. In Table 1, recent research works based on the above-discussed technologies have been reviewed.

3.2. Far Field (Radiative) WPT System

Far-field or radiative WPT system are applicable for long-range applications. The transmitters of such systems must be highly directive due to radiative losses. There are mainly two kinds of radiative power transmission: directive and non-directive. Radio waves, microwaves, and laser beams are the methodology mostly used in this regard.

3.2.1. Applications

Far-field WPT systems can be used in a variety of situations such as low-power sensor networks, where the whole system can be impinged with microwave radiation at low power levels without exceeding safety standards. High power uses for far-field WPT systems include military, industrial or space where the benefit of the ability to use WPT far outweighs the overall cost of the system itself.

3.2.2. Methodology

In such wireless power transfer systems, power is transmitted from a transmitting antenna to a receiving antenna where it is successively filtered, rectified into DC power, and supplied to the load. The major problems involved in this methodology include transmission losses and rectification losses.

For a given transmitter with a constant transmit power, the received power is given by the Friis Equation [126]:

$$P_r={\left(\frac{1}{4\pi Rf}\right)}^2{G}_t{G}_r{P}_t,$$

(12)

where $R,f,G_t,G_r,P_t,P_r$ are the distance between the transmitter and receiver, frequency of transmission, transmitter antenna gain, receiver antenna gain, transmitted power and received power, respectively.

As can be seen, the received power is inversely proportional to the square of the distance. Similarly, as frequency increases, the free path losses also increase and the received power decreases. Thus, to achieve efficient power transfer, care must be taken to design antennas with high directivity so as to point the radiated power towards the receiver and high rectifier efficiency. Further, the transmitted power can also be increased so as that the power received is at a desirable level.

3.2.3. High Directivity Antennas

An easy way to increase the directivity of the system is to use transmit antennas that are inherently directional, such as parabolic, helical, or YagiUda antennas. These antennas can be pointed directly at the receiver so as to concentrate the radiated power on the receiver antenna. Another common method for increasing the directivity of an antenna is using multiple antennas in an array. The more the number of antennas that are used in the array, the higher the directivity that is achieved [127]. High directivity dipole antenna arrays can also be achieved by designing the antenna such that all the elements of the array radiate in phase [128]. When used as a transmitter, antenna arrays can be used to control the effective area they cover. This is called beam-forming and can be used to target a moving receiver such as satellites, or drones [129].

The received power of an antenna can be increased by placing it at the focal point of a parabolic dish reflector (PDR). The PDR converts the far-field planar waves into spherical waves by focusing them at its focal point. This increases the electric field density, and thus the received power [130].

Each diode shown in Figure 9 leads to some energy loss due to the conduction resistance and the diode voltage drop. Due to this, Schottky diodes must be used. Schottky diodes are made to have low forward voltage drops and fast recovery times and are ideal for rectification at higher frequencies.

Table 2. Comparison between different rectifier topologies.

Property	Series [131]	Greinacher [132]	Voltage Doubler [133]
RF Bands(GHz)	0.2–0.5	0.9, 2.45	0.5, 0.9, 1.8, 2.45
Power Efficiency	15–20dBm	4% (For each Freq.)	15% (For each Freq.)

shows a comparison amongst a few rectifier topologies.

More recently the use of CMOS rectifiers has been considered [134]. CMOS devices have lower conduction resistance. This leads to lower conduction losses at lower frequencies. At higher frequencies, parasitic capacitance dominate and contribute to the overall loss of the system. Further, CMOS devices are active and need a gate drive voltage to turn them on. A part of the received power must be fed into the gates of these devices to turn them on. Due to this, such devices are only viable if the efficiency gained from the lower conduction resistance outweighs the power loss due to feeding the signal to the gates.

3.2.4. Maximum Power Point Tracking

Maximum Power Point Tracking (MPPT) is widely used in harvesting energy from renewable resources, such as solar and wind. This method controls the amount of load that the generator sees in order to achieve maximum power transfer.

Solar cells in particular can behave differently depending on the time of day and amount on sunlight. Due to this, an ideal operating point exists where maximum power can be harvested from the panels [135]. Solar energy can be thought of as a form of high-frequency signal which is received by the photo voltaic panels and converted into usable energy. The principles used to harvest this high-frequency radiation might also be used fully in converting lower-frequency radiation into usable energy.

Wireless power transfer is old and revolutionary technology. Near-field wireless power transfer methods are used widely for charging phones and body area networks. Wireless power transfer reduces the wear on normal connectors and reduces a common mechanical failure point. Far-field wireless power transfer can be used for powering remote sensor networks that do not require much power. Far-field power transfer can also be used in areas where transmitting harmful radiation is not concern such as the military and space.

4. An Insight for Direction of Arrival Subspace-Based Algorithms along with Improved Coherent Signal Detection Technique

In the last few decades, many researchers emphasized finding an effective signal processing-based algorithm to estimate the direction of the incident signal and mitigate the limitation associated with each algorithm. As a result, several algorithms have been developed in the last few decades, as shown in Figure 10. The conventional beamforming method was one of the earliest methods proposed by Bartlett in 1965 [136,137], and the minimum variance distortionless response (MVDR) beamforming is an improved beamforming technique that can minimize the role of the interference [138,139,140]. However, those beamforming methods were not able to separate the incident signals impinging within the same beam width [137]. Therefore, a new era of DoA has been established based on linear algebra exploiting the spatial spectrum of the incident signals [141]. The first step in analyzing the spatial spectrum includes estimating the signal and noise subspace by decomposing the array receiving signal covariance matrix into its pertained eigenstructure [142]. Based on the linear algebra theory, the k eigenvector related to the largest eigenvalue is a linear combination of all signals’ steering vectors. Those special vectors span the signal subspace and contain all the signal information. On the other hand, the remaining eigenvectors that correspond to the lowest eigenvalues span the noise subspace [142,143,144,145,146,147]. Therefore, several subspace-based methods have been developed, such as, multiple signal classification (MUSIC) [60,136,141,142,148,149,150,151] that was proposed by Schmidt and their colleagues in 1979 [136,141], improved MUSIC [141,152,153], estimation of signal parameters via rotational invariance techniques (ESPRIT) [143,151,154,155,156,157,158,159,160,161,162,163], and MI (multi-invariance)-ESPRIT [162]. Although all the aforementioned techniques are subspace-based techniques, the ESPRIT algorithm has an advantage over the MUSIC algorithm in terms of computational speed and real-time processing by utilizing the rotational invariance of the covariance matrix to estimate the DoA [155,156]. Additionally, these subspace methods have higher resolutions to estimate the DoA estimation than the beamforming method by utilizing the full rank structure deduced from the eigenvalue decomposition (EVD) component of the receiving signal covariance matrix. In the case of coherent incident signals, however, both MUSIC and ESPRIT fail as the covariance matrix rank drops to the number of uncorrelated signals. Thus, the coherence signal is not able to be distinguished in a subspace decomposition. In order to address the deleterious effect of coherent signals, a singular value decomposition ESPRIT (SVD-ESPRIT) has also been developed [144,145,155] that is capable of successfully estimating the coherent signal in DoA. As this method is a dimensionality reduction algorithm, the newly constructed matrix in this method has a smaller dimension than the original one, and it may not work for low SNR environments [155,156].

Hence, to overcome this challenge, an improved algorithm based on the Toeplitz matrix (T-ESPRIT) [156,158,159] has been introduced to retrieve the original Toeplitz nature of the signal and noise subspace matrices that could significantly be affected due to interference, multipath propagation, and the low number of snapshots. The T-ESPRIT method can be used to reconstruct a matrix by obtaining the correlation between the first element of the maximum eigenvector, and the remaining elements of that eigenvector [156,158,159], and then rearranging it to be a Hermitian Toeplitz matrix. Accordingly, a singular value decomposition (SVD) or eigenvalue decomposition (EVD) can be used to get a new full-rank signal or noise subspace. The new signal or noise subspace can be used later on in the conventional ESPRIT or MUSIC algorithms to estimate the DoA. Finally, another method based on modified singular value decomposition (MSVD-ESPRIT) [144,145,155] has been introduced as another means to improve the resolution of the DoA under a low SNR environment. The MSVD technique is based on the idea of forward-backward spatial smoothing (FBSS) [154,155,164,165], dividing the original N size array by the number of forward and backward subarrays of size P, and then reconstructing a new matrix by an exchange anti-diagonal matrix. The new reconstructed matrix will then pass through SVD or EVD to get the new signal and noise subspaces used later in the conventional ESPRIT or MUSIC algorithms to estimate the DoA.

4.1. Algorithms Methodology

4.1.1. Minimum Variance Distortionless Response (MVDR) Beamforming

Minimum variance distortionless response (MVDR) beamforming is an improved beamforming technique based on creating a steerable beam that will steer toward the directions of the signal of interest (SOI) while nulling the other undesired signals or signals of not interest (SONI) [138]. Thus, the MVDR can significantly minimize the role of the undesired interference and ultimately improve the array output SNR. The central hypothesis of the MVDR relies on reducing the power in the output while keeping the gain constant in the direction of SOI, usually unity [139], by applying the below constrain: [138,139,140],

$${E\left[\right|y\left(\theta \right)|}^2]=min{W}^H{R}_{x}W,\mathrm{subjected}\mathrm{to}{W}^{H}A\left({\theta }_{\circ }\right)=1,$$

(13)

where $E\left[\right|y\left(\theta \right){|}^2]$ is the output power, W is the element weight, $R_x$ is received signal covariance matrix, and $A\left({\theta }_{\circ }\right)$ is SOI manifold. Then the required array element weights can be estimated as [138,139,140]:

$$W=\frac{R_x^{-1}A\left(\theta \right)}{A^H\left(\theta \right)R_x^{-1}A\left(\theta \right)}$$

(14)

Finlay, the power of the MVDR algorithm can be obtained as [138,139,140]:

$$P_{MVDR}\left(\theta \right)=\frac{1}{A^H\left(\theta \right)R_x^{-1}A\left(\theta \right)}$$

(15)

4.1.2. MUSIC Algorithm

The multiple signal classification (MUSIC) algorithm is a noise subspace-based algorithm, and the MUSIC concept was first developed by Schmidt and their colleagues in 1979 [136,141]. This method of estimating the direction of arrival (DoA) is based on utilizing the orthogonality between the noise subspace and signal manifold vector [136,141]. Therefore, to represent the MUSIC algorithm, we can apply a singular value decomposing process (SVD) to extract the decomposition characteristic for the covariance matrix of a certain array, resulting in two orthogonal signals and noise subspaces. Then, based on the fact that the incident signals are orthogonal to the noise subspace, a spectrum function associated with noise subspaces and searching angles may be employed to find the peak power that will indicate the DoA of the signal [136].

To model MUSIC algorithm, we can consider a uniformly excited linear array (ULA) as shown in Figure 11 with M number of sensors and N number of incident signals impinging the ULA at angles ${\theta }_1$, ${\theta }_2$, ..., ${\theta }_N$.

The array output vector at time t is then given by [60,136,141,142,148]:

$$X\left(t\right)=AS\left(t\right)+N\left(t\right)$$

(16)

where

$$X\left(t\right)={[x_1\left(t\right),x_2\left(t\right).\dots \dots .x_M\left(t\right)]}^T$$

(17)

$$S\left(t\right)={[s_1\left(t\right),s_2\left(t\right).\dots \dots .s_N\left(t\right)]}^T$$

(18)

$$N\left(t\right)={[n_1\left(t\right),n_2\left(t\right).\dots \dots .n_M\left(t\right)]}^T$$

(19)

$$A=[a\left({\theta }_1\right),a\left({\theta }_2\right),\dots \dots ,a\left({\theta }_N\right)]$$

(20)

$$a\left({\theta }_n\right)={[1,e^{(-j(2\pi /\lambda )dsin\left({\theta }_n\right))}...,e^{(-j(M-1)(2\pi /\lambda )dsin\left({\theta }_n\right))}]}^T$$

(21)

where $X_M\left(t\right)$, ($m=1,2...,M$) is the input of the $m^{th}$ element, $S_n\left(t\right)$ and $a\left({\theta }_n\right)$ are the signal complex amplitude, steering manifold vector, respectively, $n_m\left(t\right)$ is the noise of the $m^{th}$ element and assumed to be uncorrelated zero-mean Gaussian white noise with the power of each entry equal to ${\sigma }_n^2$, $\lambda$ is the wavelength of the signal, d is the inter-spacing between array elements and ${(.)}^T$ is the transpose operator. $\theta$ is elevation angle of incident signal, $\tau$ is signal arrival delay between consecutive array elements. Then the covariance matrix ($R_x$) can be evaluated as below [60,136,141,142,148]:

$$R_x=\frac{1}{K}\sum _{i=1}^{K}x\left(t\right)x^H\left(t\right)$$

(22)

$$\Rightarrow R_x=E\left[X\left(t\right)X^H\left(t\right)\right]=E\left[(AS+N)\right]{(AS+N)}^H]$$

(23)

where K is number of snapshots, E[] is the mean or expectation of two random variables and then $R_x$ can be expressed as [60,136,141,142,148]:

$$R_x=AE\left[S{S}^H\right]A^H+E\left[N{N}^H\right]=A{R}_s{A}^H+{\sigma }^{2}I$$

(24)

where $R_s$ is the source auto-covariance matrix and ${\sigma }^2$ is the assumed white Gaussian noise variance. The eigenvalue decomposition for the covariance matrix is as [60,136,141,142,148]:

$$R_x=A{R}_s{A}^H+{\sigma }^{2}I=U_s{\Lambda }_s{U}_s^H+U_n{\Lambda }_n{U}_n^H$$

(25)

where $U_s$ is the signal subspace corresponding to larger eigenvalues, and $U_n$ is the nose subspace corresponding to the smallest eigenvalues, ${\Lambda }_s$, and ${\Lambda }_N$ are the diagonal matrix for the pertained signal and noise eigenvalues, respectively. Let ${\lambda }_i$ be the $i^{th}$ eigenvalues of the matrix Rx, and $v_i$ is eigenvector corresponding to ${\lambda }_i$, then [136]:

$$R_x{v}_i={\lambda }_{i}vi$$

(26)

Let ${\lambda }_i$= ${\sigma }^2$ be the minimum of $R_x$, then $v_i$ will be eigenvector corresponding to the noise,

$$R_x{v}_i={\sigma }^2{v}_i$$

(27)

by substituting (24) in (27), we get [136]:

$${\sigma }^{2}vi=(A{R}_S{A}^H+{\sigma }^{2}I)v_i$$

(28)

and expanding the right polynomial and then comparing it with the left term, we get [136]:

$$A{R}_S{A}^H{v}_i=0,$$

(29)

$$\Rightarrow A^H{v}_i=0,i=N+1,N+2,\dots ,M.$$

(30)

The above equation indicates that the noise eigenvector $\left(v_i\right)$, which corresponds to the noise eigenvalue, is perpendicular to the column vector of the matrix A [136], and since each row of A corresponds to the direction of signal sources thus the noise eigenvectors and the steering vectors that makeup matrix A will be used to form the MUSIC angular spectrum which is given by [60,136,141,142,148]:

$$P_{\mathrm{MUSIC}}=1/\left(a^H\left(\theta \right)U_n{U}_n^{H}a\left(\theta \right)\right).$$

(31)

The noise eigenvector can be constructed by first separating the signal subspace eigenvalue and eigenvector, which is equal to the number of signals N; then, by taking the rest of the $(M-N)$ eigenvalues and eigenvectors as noise subspace. Finally, the noise matrix $U_n$ can be obtained as shown below [136,142]:

$$U_n=[U_{N+1},U_{N+2},\dots ,U_M].$$

(32)

4.1.3. Conventional ESPRIT Algorithm (LS-ESPRIT)

Estimation of signal parameters via rotational invariance techniques (ESPRIT) is another type of subspace-based method to estimate the direction of arrival (DoA). Contrary to the MUSIC algorithm, this method depends on signal subspace rather than the noise subspace, and it does not involve an exhaustive search through all possible steering vectors to estimate DOA. Thus, it dramatically reduces the computational and storage requirements compared to the MUSIC algorithm [151]. The central hypothesis of the conventional method or least square ESPRIT (LS-ESPRIT) is by dividing the M element array into two overlapping sub-arrays with length $(M-1)$ as shown in Figure 12 [155,156,157] and then utilizing the characteristic of the rotational invariance between these two sub-arrays [155,156,160]. Then the received impinging signal on each sub-array will be [151,155,156,157]:

$$X_1\left(t\right)=AS\left(t\right)+N_1\left(t\right)$$

(33)

$$X_2\left(t\right)=\overline{A}S\left(t\right)+N_2\left(t\right)$$

(34)

where $\overline{A}$= A*$\psi$, and $\psi$ represents the manifold difference between two arrays as illustrated below:

$$\psi =\mathrm{diag}[e^{-jkdsin\left({\theta }_1\right)}{e}^{-jkdsin\left({\theta }_2\right)}..,e^{-jkdsin\left({\theta }_N\right)}]$$

(35)

where k is the wave constant ($k=2\pi /\lambda$). Then by applying EVD for the covariance matrices for both sub-array signals ($X_1$ and $X_2$), we can deduce the signal subspaces for the two sub-arrays as $V_1$ and $V_2$, respectively, which can be represented as $[\overrightarrow{e_1},\overrightarrow{e_2},...,\overrightarrow{e_k}]$. Now, the larger N eigenvectors related to larger N eigenvalues span the signal subspace, while the other $M-N$ eigenvectors that are related to $(M-N)$ eigenvalues span the noise subspace, and they are orthogonal to the signal subspace [166].

By conducting the EVD for both subarrays $X_1$ and $X_2$, we get their signal subspace $U_{s1}$ and $U_{s2}$ as shown below [157]:

$$R_{x1}=E\left[X_1{X}_1^H\right]=A{R}_s{A}^H+{\sigma }^{2}I=U_{s1}{\Lambda }_s{U}_{s1}^H+U_{n1}{\Lambda }_n{U}_{n1}^H$$

(36)

$$R_{x2}=E\left[X_2{X}_2^H\right]=\overline{A}{R}_s{\overline{A}}^H+{\sigma }^{2}I=U_{s2}{\Lambda }_s{U}_{s2}^H+U_{n2}{\Lambda }_n{U}_{n2}^H$$

(37)

Since the two subarrays are translationally related, their signal subspaces will be related by a unique non-singular transformation matrix $\mathsf{\Phi }$ such that [151]:

$$U_{s1}\mathsf{\Phi }=U_{s2}.$$

(38)

Since the signal subspace$\left(U\right)$ is equal to the subspace of the manifold $\left(A\right)$, the $\mathrm{span}\left(U\right)=\mathrm{span}\left(A\right)$ and there must also exists a unique non-singular transformation matrix T so that [144,151,166]:

$$U_{s1}=AT$$

(39)

$$U_{s2}=\overline{A}T$$

(40)

$$U_{s2}=A\psi T$$

(41)

and finally we can derive from (38) and (41) that:

$$U_{s1}\mathsf{\Phi }=A\psi T,$$

(42)

and by substituting (39) in (42):

$$AT\mathsf{\Phi }=A\psi T,$$

(43)

and next, multiplying both side by $A{T}^{-1}$, we get [145,151,160,164]:

$$T\mathsf{\Phi }{T}^{-1}=\psi ,$$

(44)

and finally from (38) we can infer that:

$$\mathsf{\Phi }=U_1^{-1}{U}_2.$$

(45)

Thus, the eigenvalues of $\mathsf{\Phi }$ = diag [${\lambda }_1$, ${\lambda }_2$, ..., ${\lambda }_N$] must be equal to the diagonal elements of $\psi$ such that ${\lambda }_1=e^{-jkdsin\left({\theta }_1\right)},{\lambda }_2=e^{-jkdsin\left({\theta }_2\right)},...,{\lambda }_N=e^{-jkdsin\left({\theta }_N\right)}$. Once the eigenvalues of $\mathsf{\Phi }$ (rotational invariance) are calculated through EVD or SVD, we can estimate the angle of arrival as below [144,151,155,160]:

$${\theta }_i={sin}^{-1}(-\angle \left({\lambda }_i\right)/kd)$$

(46)

where ∠ represents angle; same as MUSIC, the traditional ESPRIT can not detect the coherence signals. Thus, a single value decomposition technique (SVD-ESPRIT) has been introduced to decorrelate the covariance matrix and retrieve the full rank matrix in order to estimate the coherent incident signals.

4.2. SVD Technique for Coherent Sources (SVD-ESPRIT and Improved MUSIC)

Even though the traditional MUSIC and ESPRIT algorithm can accurately estimate the DoA of non-correlated impinging signals, the full rank (N) cross-spectrum matrix ($R_x$) is required to resolve the coherent sources successfully. Based on that, the traditional method fails to estimate the coherent signals. In certain scenarios, the number of eigenvalues estimated from the decomposition covariance matrix (SVD or EVD) will reduce to r (number of incoherent sources). Therefore, the remaining $(N-r)$ (coherent signals) eigenvalues will not be considered in subspace decomposition [150]. Thus, the signal subspace dimension turns out to be less than the rank of the manifold array, A.

For instance, if all the impinging signals are coherent, only one eigenvalue will be larger than the noise variance, i.e., ${\lambda }_1>{\lambda }_2={\lambda }_3=.....{\lambda }_N={\sigma }^2$. Therefore, ${\Lambda }_s$ in this case will be a scalar and equal to (${\lambda }_1$) while $U_s$ will be a vector and equal to $\left(u_1\right)$. Consequently, the matrix $U_s{\Lambda }_s{U}_s^H$ can be rewritten as ${\lambda }_1{U}_1{U}_1^H$ and the rank of this matrix drops to 1. At the same time, the dimension of the noise subspace expands to $(M-1)$, which yields that the steering vector $a_n$ pertained to coherent signals will no longer be upright to the noise subspace, and that leads the traditional MUSIC and ESPRIT to fail to detect the DoA of the coherent sources [145]. Accordingly, the SVD algorithm has been developed based on the reconstruction of matrices to restore the full rank matrix property of the signal subspace or noise subspace of the coherent signal sources.

To module a modified algorithm, let us assume that a total of N numbers of plane waves impinge on an M element array antenna. The rank of the array manifold is N, and the signal covariance matrix rank is K (where $K\le N$). Now suppose the noise covariance matrix ($R_N$) is a full rank matrix as below [144,155]:

$$R_N{e}_k=\sum _{n=1}^N{\alpha }_k\left(n\right)a\left({\theta }_n\right)$$

(47)

where $e_k(1\le k\le K)$ is an eigenvector of the received signal covariance matrix, and ${\alpha }_k\left(n\right)$ is a linear combination factor. If the noise is white Gaussian noise, its covariance matrix will be an identity matrix and then the above equation will simplify to [144,145]:

$$e_k=\sum _{n=1}^N{\alpha }_k\left(n\right)a\left({\theta }_n\right)$$

(48)

This equation indicates that whether the incident signals are coherent or not, the eigenvector related to the largest eigenvalue is a linear combination of all such incident signals’ steering vectors. Now, when $K=1$ (i.e., the signal sources are completely coherent), the equation will reduce to [144,145]:

$$e_1=\sum _{n=1}^N{\alpha }_1\left(n\right)a\left({\theta }_n\right)=AT$$

(49)

where $e_1={[e_{11},e_{12}...,e_{1M}]}^T$ is the first eigenvector column in the signal subspace of the covariance matrix and $T=[t_1,t_2...,t_N]$ is a linear combination factor. From the above equation, we can conclude that the largest eigenvector is a linear combination of all incident signals’ steering vectors and contain all the required signal information [144,145].

Now, the vector $e_1$ can be used to re-construct a new covariance matrix $R_1$, in the same manner the vector $e_2$ (the first eigenvector column in the signal subspace of the second subarray $\left(U_2\right)$) can be used to re-construct a new covariance matrix $R_2$ as shown below [144,145,155]:

$$R_1=\left[\begin{array}{cccc}{e}_{11}& e_{12}& ...& e_{1p}\\ e_{12}& e_{13}& ...& e_{1p+1}\\ ...& ...& ...& ...\\ e_{1m}& e_{1m+}& ...& e_{1M}\end{array}\right],$$

(50)

$$R_2=\left[\begin{array}{cccc}{e}_{21}& e_{22}& ...& e_{2p}\\ e_{22}& e_{23}& ...& e_{2p+1}\\ ...& ...& ...& ...\\ e_{2m}& e_{2m+}& ...& e_{2M}\end{array}\right],$$

(51)

where the number of sensors per subarray $m>N$, number of subarrays $P>N$, and $m+p-1=M$. Hence, we can infer that rank $[R_1$ or $R_2]>N$. Hence, we can consider our array as a single array with only one new constructed matrix $\left(R1\right)$, or two overlapping arrays with two new constricted matrices $(R1$ and $R2)$. Accordingly, we can apply the aforementioned conventional MUSIC or ESPRIT algorithms to estimate the coherent signal in the direction of arrival. Though this SVD technique can decorrelate the coherent signal, it will not work rigorously under a low SNR environment. As a result, several modified algorithms have been developed to improve the algorithm performance, namely, the Toeplitz matrix and forward-backward special smoothing.

4.3. Algorithm Improvement Techniques

4.3.1. Toeplitz Matrix Technique (T-ESPRIT)

As mentioned earlier, the single value decomposition algorithm (SVD algorithm) can estimate the coherent signal DoA [144,155,156]. However, it still will not work well under a low SNR environment. Therefore, to obtain a higher resolution DoA, the Toeplitz matrix reconstruction technique has been introduced. Consequently, referring to the early calculated covariance matrix in (24) and the received impinging signal on each sub-array of ESPRIT in (33) and (34), we can integrate the two sub-arrays receiving signals in a single matrix with ESPRIT algorithm to obtain the relationship between them [155,156]:

$$X=\left[\begin{array}{c}{X}_1\\ X_2\end{array}\right]=\left[\begin{array}{c}{A}_1\\ A_1\psi \end{array}\right]S+\left[\begin{array}{c}{N}_1\\ N_2\end{array}\right]=A^{{}^{\prime }}S+N.$$

(52)

Then the covariance matrix for the ESPRIT can be deduced as:

$$R_x=E\left[X{X}^H\right]=A^{{}^{\prime }}{R}_s{A}^{{}^{\prime }H}+{\sigma }^{2}I,$$

(53)

where $R_S=E\left[S{S}^H\right]$. Then we process EVD over the covariance matrix $R_x$:

$$\mathrm{EVD}\left(R_x\right)=U_s{\Lambda }_s{U}_s^H+U_n{\Lambda }_n{U}_n^H.$$

(54)

As shown early in the modified algorithm, the eigenvector related to the largest eigenvalue is a linear combination of all incident signals’ steering vectors; the two largest eigenvectors $e_1$ and $e_2$ corresponding to the largest two eigenvalues have thus been chosen to achieve the decoherence, where:

$$e_1=[e_{11},e_{12}...,e_{1M}],$$

(55)

$$e_2=[e_{21},e_{22}...,e_{2M}].$$

(56)

We take $e_{11}$ as a reference and conduct the correlation function between $e_1$=[$e_{11},e_{12}...,e_{1M}$] & $e_{11}$, respectively. Then we construct the vectors below [156]:

$$R(m-1)=E\left[e_{11}{e}_{1m}^H\right]$$

(57)

where $m=1,2..,M$, and by changing the m from 1 to M, we obtain [156]:

$$\left[r\right(0),r(1),\dots ..,r(M-1\left)\right],$$

(58)

which satisfies [156]:

$$[r\left(0\right),r\left(1\right),\dots ..,r(M-1)]=[E\left[e_{11}{e}_{11}^H\right],\dots ..,E\left[e_{11}{e}_{1M}^H\right]].$$

(59)

Additionally, [156]:

$$R(-m+1)=E\left[e_{1m}{e}_{11}^H\right],m=1,2..,M.$$

(60)

From above constructed vectors and in order to have better DoA resolution under low SNR, a new matrix $Y_1$ can be constructed as shown below [156,158,159]:

$$Y_1=\left[\begin{array}{cccc}r\left(0\right)& r\left(1\right)& ...& r(M-1)\\ r(-1)& r\left(0\right)& ...& r(M-2)\\ ...& ...& ...& ...\\ r(-M+1)& r(-M+2)& ...& r\left(0\right)\end{array}\right],$$

(61)

where $r(-m+1)=r(m-1)$, and similarly we can get $Y_2$ from $e_2$. The constructed $Y_1$ and $Y_2$ matrices are both Hermitian Toeplitz matrices [156]. Then, singular value decomposition could be applied on Hermitian Toeplitz matrices $Y_1$ and $Y_2$ to get the modified signal subspace or noise subspace which would be used later in ESPRIT or MUSIC algorithms, respectively.

4.3.2. Forward/Backward Spatial Smoothing Technique (FBSS) (MSVD-ESPRIT)

The central hypothesis of the spatial smoothing method is dividing the ULA into the several overlapping forward and backward sub-arrays, each of size P, as shown in Figure 13 to have an improved resolution to estimate the coherent incident signals DoAs. To model FBSS, a new matrices have been constructed referred as $Q_1$ and $Q_2$ with the idea of forward/backward spatial smoothing technique as illustrated below [144,145,154,155,161,163,167,168,169,170,170]:

$$Q_1=1/2(R_1+J_{mxm}{R}_1^{\ast }{J}_{pxp}),$$

(62)

$$Q_2=1/2(R_2+J_{mxm}{R}_2^{\ast }{J}_{pxp}),$$

(63)

$$J=\left[\begin{array}{cccc}0& 0& ...& 1\\ 0& 0& ...& 0\\ ...& ...& ...& ...\\ 1& 0& ...& 0\end{array}\right],$$

(64)

where $R_1$ and $R_2$ are the improved reconstructed covariance matrix and J is an exchange matrix whose anti-diagonal elements are equal to 1 and all other elements are equal to 0. Then a singular value decomposition (SVD) is performed on the modified matrices ($Q_1$ and $Q_2$) to get the modified signal subspaces $U_{S1}$ and $U_{S2}$ that can be processed later through the ESPRIT algorithm. Alternatively, an eigenvalue decomposition can be performed for $Q_1$ to get the noise subspace that can be processed later through the MUSIC algorithm.

5. Wireless Body Area Networks and Its Energy Harvesting Techniques for Health Care

The continuous advancement and increasing interest in wireless communication technologies and the extensive use of wireless sensors have accelerated the growth of the Wireless Body Area Network (WBAN). A combination of both tiny sensors and wireless communication technology has made the wireless body area network (WBAN) one of the most promising fields where these sensors can be attached either to the body or even implanted under the skin. Using the WBAN, the patient experiences greater physical mobility as he is no longer obliged to stay in the hospital.

Since nowadays people are more concerned about their health, it has introduced immense pressure for administration, healthcare contributors, and healthcare diligence to develop the quality and quantity of the overall health care system. Previously, patients needed to carry their health records with them so that doctors can easily analyze which medicines are working well and can figure out any drawbacks caused by any of the prescribed medicines which is a cumbersome task to do all the time [171]. Therefore, it gave arise the need for a technology that can remotely monitor the health and share information with care providers or hospitals. Electronic health monitoring and fitness monitoring, online consultation with specialists, diagnosis, and remote healthcare are possible these days through exploiting various communication technologies. Among all of these helpful ways, the wireless body area network (WBAN) has become the most convenient, cost-effective, and accurate technology since it enables real-time and continuous health monitoring in the healthcare domain [172,173]. Here, is the general structure of a WBAN shown in Figure 14:

WBAN is based on frequency-based wireless networking technology and acts as a sub-part of the Wireless Sensor Network (WSN) [174]. Based on the operating environment, a Wireless Body Area Network (WBAN) connects lightweight independent nodes such as sensors and actuators that are situated either in the clothes, on the body, or implanted under the skin of a person. The WBAN is formally defined by IEEE 802.15 as, “a communication standard, optimized for low power devices network that typically expands over the whole human body and the nodes are connected through a wireless communication channel [175]. These intelligent and small biomedical sensors then collect, process the information, and finally send it to a medical server for further analysis.

As WBANs provide easy and cost-effective solutions in healthcare based on their different monitoring support, it is getting more hits with time and more development is required in this field. However, new paradigms and protocols in the proximity of the human body may challenge its design issues [176]. On the other hand, the increased usage of WBAN in e healthcare needs more emphasis on its security. It must ensure the integrity and confidentiality of the sensed data [177]. WBAN in combination with the mobile phone can work as a gateway to maintain the network. Several energy harvesting mechanisms are being discussed in [178] as increasing the battery life can be the biggest challenge in handling the overall system.

The architecture of the paper is the following parts: describes the applications of in WBAN, design considerations, devices used in WBAN, technologies enabling the WBAN, different energy harvesting techniques and finally s describes the conclusion and future scope.

5.1. Application of a WBANs

We can categorize the usage models of WBANs based on their wide range of monitoring applications in the fields of healthcare, emergency, military, sports, interactive gaming, and many others [179]. However, initially, BANs are expected to appear in the healthcare domain, especially for continuous monitoring and analyzing vital parameters of patients suffering from chronic diseases such as diabetes, asthma, heart attacks, stress, and physical rehabilitation along with a wide range of real-time monitoring of necessary health parameters including blood pressure, electroencephalogram (EEG), electrocardiogram (ECG), carotid pulse, glucose rate, body temperature.

5.2. Medical Application

5.2.1. Cancer Application

The main usage of WBAN is our medical sector. Nowadays millions of people are getting affected by cancer and the number of affected people is increasing every year [180]. A set of small sensor nodes have capabilities to detect nitric oxide which is usually produced by cancer cells. Sensor nodes can be implanted in the infected locations in the patient body and thus allow medical professionals to detect cancer tumors without any biopsy.The ultra-wideband antennas are also used in the treatment of breast cancer with reduced SAR (Specific Absorption Rate) [181].

5.2.2. Cardiovascular Application

In WBAN, cardiovascular applications are used to monitor cardiovascular disease which is one of the primary reasons for death since more than twenty million people have been affected by this disease in the world [182].

5.2.3. Diabetes Application

Since uncontrolled sugar levels may cause other serious complications in the body like blindness, stroke, kidney disease, heart disease, and high blood pressure [183]. WBANs can provide effective treatment to diabetes patients by providing continuous and accurate monitoring of glucose levels in the blood.

5.2.4. Stress Application

Chronic stress can lead a man to other severe diseases like a heart attack, stroke, and morality. With continuous monitoring, WBANs have made it possible to ensure proper treatment to the patients [184].

5.2.5. Remote Monitoring

With the seamless connectivity of the internet, biosensors can continuously or periodically read heart rate, body temperature, pulse rate, respiration rate, blood pressure, and other important physiological parameter the patients’ health condition and send data to the doctors [185].

5.2.6. Fitness, Performance, and Well-Being Tracking

Normally athletes and military personnel require these kind of services whose main target is to improve stamina. Previously, gym instructors used to keep the records of their trainees for further analysis. However, now this work is much easier with the contribution of WBANs. For example, the Tomtom watch can show the calorie burnt after a certain time of jogging and exercise [185]. Recently, corporate offices are analyzing the sensory data to keep their employees active and productive. For example, Hitachi’s business microscope monitors are used to analyze vital parameters such as movements, ambient lighting, voice level to improve their work environment [186].

5.3. Nonmedical Application

5.3.1. Lifestyle and Entertainment Application

WBAN plays a vital role in our daily life. Virtual gaming purposes, personal item tracking, exchanging digital profiles or business cards, and consumer electronics are also some of the popular usages of WBAN [187]. Moreover, navigation support while walking, driving, exploring a new areas is the most basic application of WBAN [188].

5.3.2. Sports

Checking up the health condition of the players during any game time has been made possible through WBANs since the devices are wearable. It can send real time such as heart rate, temperature, blood pressure, activity, and posture of any athlete in sports [172].

5.3.3. Emergency

The WBAN sensors has the ability to sense any abnormal situation and send a notification to the alarm thus helping the patients. Thus, it is playing a vital role in sensing any serious situation and saving valuable lives in both household and industries [185].

5.3.4. Military Purpose

WBAN can provide potential benefits like connectivity, survivability for military networks, and applications regarding the health (heart rate, blood pressure, hydration level, etc.) of soldiers and field personnel during a mission. On the battlefield, the soldiers can communicate with each other about giving commands to attack, run, and share information [189].

5.4. Characteristics of WBAN

5.4.1. Difference between WBAN and WSN

Although the main target of both a general Wireless sensor network (WSN) and a Wireless Body Area Network (WBAN) is to sense and send data wirelessly, there are some technical differences in the case of its implementations and use cases. Basically, WBAN is a type of WSN, where, to enable device-to-device communication, useful techniques from Wireless Sensor Network and ad hoc networks may be used. However, A WSN may not face the same problems encountered as in a human body monitoring system and both of the applications have different requirements and challenges. For example, the WBAN may possess a small-scale deployment with a number of around 20 sensors by a few persons whereas the WSN normally consists of hundreds of nodes deployed in an environment with a range of several kilometers. In addition to these differences in terms of the scale of the network and the number of sensors, there are also some general differences between WSN and WBAN in terms of accuracy, mobility, node replacement as well as energy scavenging power as shown in

Table 3. The general differences between WBAN and WSNE.

	WSN	WBAN
Scale	Large scale as ranges with several kilometers.	Small scale limited by the human body.
Node number	A small number of nodes required.	A large number of nodes.
Node size	A miniaturized node is not mandatory.	A miniaturized node is a must as implemented on human body.
Node replacement	Easier to handle.	Difficult to replace and sometimes impossible in the case of implanted devices.
Data rate	WSN is employed for irregular event-based monitoring.	WBAN may occur in a more periodic manner and show a stable data rate.
Accuracy	Compensated by redundancy.	Transmitted data must be accurate.
Mobility	Nodes are stationary.	Nodes may move as the body is not stationary.
Latency	Better latency as battery replacement is much easier.	Need to maximize battery life for higher latency.
Energy scavenging power	Wind, solar power	Body movement and temperature
Biocompatibility	Not considered	Is a must

elsewhere mentioned in [190,191,192].

5.4.2. General Architecture of WBAN

This section provides an outline of the general architecture of a WBAN. Since, in the last decade, the wireless communication technologies and standards have extended exponentially, the WBAN technology results as a revolutionary outcome of the existing WSN technology where each type of the network can be supplemented by several wireless technologies such as IEEE 802.15.1 (Bluetooth) or 802.15.4 (Zigbee), IEEE 802.11 (Wi-Fi), and WMAN IEEE 802.16 (WiMax) [193]. A WBAN generally comprises multiple sensors where sensors sample, process, and communicate vital health parameters like heartbeat rate, vascular blood pressure, and/or blood oxygen saturation and provide real-time feedback to the user and medical personnel.

Generally, we can distinguish WBAN entities into two major categories, sensor node and sink or gateway node where the former entity is responsible for data collection from the human body through sensors while the latter entity sends it to other servers and communication networks.

WBAN communication architecture is classified into three categories such as intra-WBAN communication (Tier-1), inter-WBAN communication (Tier-2), and beyond-WBAN communication (Tier-3) [194]. Figure 15 shows the communication tiers in an active typical WBAN, where the devices are distributed all over the body within the central network architecture. Moreover, the location of such a device depends on the specific application and movement of the body [195,196].

• First-tier of WBAN architecture is realized by body sensor units that are placed either outside or inside of the human body. These sensors detect the physiological data signals from the human body, convert it to digital forms, and then transmit through wireless from the human body and finally send it wirelessly to the second tier. This communication is also known as intra-BAN communication.
• The Second-tier builds a connection between the first-tier and third-tier wirelessly. The second tier is also known as inter-BAN communication and consists of personal server units. These units get data from sensors, process it, and format the processed results to convey to the upper, third-tier if necessary.
• Third-tier comprises of user machines, where end users are data experts who involves decision making based on the data from tier-2 such as sending some caretaker or ambulance to the patient, taking some specific diet for the sportsman.

5.5. Technologies Enabling WBAN

5.5.1. Bluetooth

Bluetooth is an IEEE 802.15.1 standard usually known as WPAN (Wireless Personal Area Network). In the early 2000s, a WBAN project named MOBI Health used Bluetooth technology to transmit sensor data from a front-end device to a mobile phone or PDA [185]. It is a wireless communication standard operating within the ISM 2.4 GHz frequency band in the range 2400–2483.5 MHz [197]. This technology was designed as a short-range wireless communication standard, anticipated to form a network with security and low power consumption. According to [198], A particular Bluetooth network forms a Pico net where a Bluetooth device works as a master and another seven Bluetooth devices. Again, more than one piconet can build another network named Scatternet. Since Bluetooth is suitable for short-distance data transmission applications, it is possible to join together numerous piconet into a large scattered and to expand the physical size of the network beyond Bluetooth’s limited. Ref. [199] Says, based on the coverage area, Bluetooth can be classified into three devices: 1. ranging from 1 to 100 m and 2. Different transmission powers ranging from 1 mW to 100 mW with 3 Mbps data rate.

5.5.2. Bluetooth Low Energy (BLE)

Bluetooth low energy has become very popular due to its low power consumption and low latency. It is the extended version of Bluetooth and it requires a low duty cycle [184]. This technology is also known as Bluetooth 4.0. According to [200], it can build star networks like Bluetooth and is designed to operate small mobile devices with a data rate of 1 Mbps.

5.5.3. Zigbee/IEEE 802.15.4

ZigBee is another popular technology for WBAN applications which is designed to work below a low power consumption environment with data ranging from 20 kbps to 250 kbps [201]. During sleep mode, a Zigbee-based device can be working without any recharge for a long time due to low power consumption [202]. However, due to its low data rate and interference with WLAN transmission sometimes, Zigbee technology is not suitable for WBAN.

5.5.4. WiFi

WiFi is an IEEE 802.11 standardized solution for Wireless Local Area Network (WLAN) which is suitable technology for video conferencing, voice calls, and video streaming that has to fulfill larger data transfer requirements. Though all smart phones, laptops and tablets are Wi-Fi integrated; high energy consumption can be an issue to handle [172].

5.6. Hardware and Devices Used in WBAN System

A WBAN architecture is composed of delicate and intelligent devices that are able to transmit their data to the medical server in order to inspect and store them. These certain devices have to be considered wisely while designing a WBAN system. In the WBAN we can distinguish 3 main types of devices shown in

Table 4. Devices used in WBAN.

	WSN	WBAN
Sensor nodes	- monitors the physiological parameters and process reports wirelessly - situated either on the body or, implanted - consists of several components such as sensor hardware, a power unit, a processor, memory, and a transmitter or transceiver - monitor important physiological parameters, i.e., BP, ECG, Pulse Rate, EEG, etc.	[205,206,207]
Actuators	- Consists of a receiver or transceiver, a power unit, memory - act as a drug delivery system and analyzes data received from a sensor node - Gives feedback to patients based on analyzed data - act as a local processing unit	[171,208]
Monitoring server	- collects the information attained by the sensors and actuators and sends it to the medical server via an external gateway (internet). - also known as Body Control Unit(BCU), body gateway or, sink - requires power unit, large memory, processor, and transceiver to perform	[171,208]

.

5.7. WBAN System Requirements

This sub-section provides a brief description of system requirements that are important to design an efficient WBAN.

5.7.1. Power Source and Patients’ Safety

Since sensor nodes use power for data processing, communication, and sensing, the WBAN has to be made power efficient. Though the sensors are battery operated in the case of implanted devices, it is difficult to replace the battery frequently. In that case, different techniques should be taken to power them remotely. Considering SAR (Specific Absorption Rate) in human health, ensuring low-powered implanted devices can be a vital consideration. As the absorption of radiation increases, the SAR also gets higher. According to [203], to ensure patient safety, the IEEE C95.1-1999 limits the averaged SAR value over 1 g of tissue to 1.6 W/Kg.

5.7.2. Data Rate

The data rate in WBAN normally depends on its application such as temperature monitoring, medical imaging applications. The data rate for various applications is shown

Table 5. Data rate for different application.

Application	Data Rate
EMG	320 Kbps
Glucose monitor	1600 bps
ECG	144 Kbps
Temperature	120 bps
Audio, medical imaging	10 Mbps
endoscope	1 Mbps
EEG	43.2 Kbps
Cochlear implant	100 Kbps

[204].

5.7.3. Size and Form Factor of Sensors

Based on the application, these criteria must be selected wisely. In general, the sensors must be tiny, flexible, and lightweight for the user’s comfort. The form factor of a sensor node can directly affect its power consumption.

5.7.4. Antenna and Surrounding Medium Specification

Since the devices used in WBAN are tiny, miniaturized antennas are mostly preferred. However, the small size of antennas may degrade their performance. In this context, a trade-off between size and performance should be considered. The human body is a dispersive medium, antenna installation inside the body might be an obstacle too [209].

5.7.5. Security

Security gets high priority while designing a WBAN which means the protection of information from unauthorized users while data is being stored and transferred. Several studies have been focused on making WBAN data transmission more secure. The data security is more because of WBAN’s infrastructure [173,210,211]. The major security requirements are shown in

Table 6. Security requirements.

Major Security Requirement	Description
Reliability	Patient-related data must be readily retrievable in case of failure of a node
Authentication	The sender must be authentic and
Accessibility	Patients’ data should be available in case of a Denial-of-Service attack (DoS)
Integrity assurance	Patient-related data must not be modified illegally during storage periods
Privacy	The data access policy must be enforced to prevent unauthorized access to patient-related data generated by the WBAN.
Non-repudiation	The source must admit the origin of every piece of patient-related data generated by it.
Accountability	If a user of the WBAN exploits their privilege to carry out forbidden actions on patient-related data, he/she should be identified and held accountable.
Key management Protocols	Trusty server, key pre-distribution, and self-imposing are the key management protocols to develop a secure application.

:

5.8. Energy Harvesting and WBAN

WBAN involves distributed, low-power, wearable, and implantable sensor nodes that can store and transmit monitored data to remote locations at a human level. These sensors usually use rechargeable batteries as their power supply but changing the power source can be burdensome, sometimes irreplaceable, and time-consuming, especially when implanted inside the human body. Thus, with the widespread use of wireless sensors, the management of their energy resources has become a topic of exploration. Recently, numerous research efforts at various layers have focused on powering the sensors via energy-efficient protocols, conservation schemes, and effective topology design. By adopting various battery recharging techniques, we can handle this issue. According to [212], these mechanisms can be classified into two in broad. They are: Wireless Power Transfer (WPT) and Energy harvesting.

5.8.1. Wireless Power Transfer (WPT)

Inductive coupling, magnetic resonant coupling, and electromagnetic radiation are some of the well-known techniques of wireless power transfer. However, due to its short-range limit, line of sight conditions in most cases static devices prefer this technique more than mobile devices.

5.8.2. Other Energy Harvesting Techniques for Wearable Devices for WBAN

Energy harvesting involves scavenging power from a variety of limitless ambient sources and converting it to usable electric energy. According to [213] it is possible to harness electricity from the atmosphere to operate electronic devices rather than rechargeable batteries. There are several potential energy harvesting sources such as body warmth, heartbeat, solar, piezoelectricity, wind, radiofrequency vibration energy, and so on [214,215,216,217]. The block diagram of the energy harvesting system is shown in Figure 16.

(a)

Ambient energy: light and vibrational energies can be good sources of energy.

(b)

AC-DC/DC-DC converter: this block is used either to convert the AC output from some harvesters into DC or from one DC to another DC level [218].

(c)

Voltage Regulators: these are used to maintain the voltage level within an acceptable range for energy storage and load stages [218].

(d)

Energy Storage: Here, the characteristics of the storage module are dependent on the specifications of the targeted application [218].

Since in a WBAN power is consumed through different operations such as real-time health monitoring, diagnosis, and intercepting, the power requirements for various wearable and biomedical devices are different. The typical power consumption of various small electronic devices is shown in

Table 7. Summary of power consumption of several wearable and medical devices.

Device	Power Requirement	Reference
Electrocardiogram (ECG) sensor	2.76 $\mathsf{\mu }$M	[219]
Pacemaker	1 $\mathsf{\mu }$W	[220]
Neural recording	1–10 mW	[187]
sensor on wristband	0.83 mW	[214]
Chest patch	0.96 mW	[214]
Spirometer	0.01 mW	[214]
Retinal prostheses	250 mV	[216]

.

5.8.3. Summary of Energy Harvesting Techniques and Their Performance

Various researches are going on to apply energy scavenging as a great energy source and it will significantly help the biomedical and implanted sensors by removing the repeated surgeries to adjust the implantable devices. Triboelectric Nano generator (TENG), piezoelectric generator (PEG) and thermoelectric generator(TEG) are the recent promising energy generators that can work either as an extension of battery or, replace the battery fully and thus prolong lifespan of the wearable biomedical devices. These EHs (Energy Harvesters) typically transduce different environmental energy sources, such as thermal, vibration, and mechanical energy into electrical energy.

According to [213], A TEG is a solid-state device that is used to convert thermal energy into electric potential, TENG is an EH device is used to transduce mechanical energy, such as respiratory movement, blood flow, walking, cardiac motion, ocean waves, and wind energy into electrical energy [221] and the fundamental mechanism of electrical energy conversion of a PEG depends on the piezoelectric material, where an external force or vibration causes a deflection on the piezoelectric structure which undergoes compression and tension, thereby generating electric potential by employing the piezoelectric effect [222].

Table 8. Different techniques and its drawbacks.

Generator Types	Working Principle	Advantages	Disadvantages	Challenges	References
Piezoelectric energy harvester	Piezoelectric effect	- Small - Fast response - No heating issues - Sensitive to strain - High output voltage - Simple excitation - Compact in size - High power density - Support well-known fabrication process	- Less efficient at low frequency - High impedance with low current - Piezoelectric materials are expensive sometimes poisonous - Require high thermal processing	- To increase the output power - Human-level frequency range - Developing power management circuit	[213,223,224]
Thermoelectric Generator	Seebeck effect	- Ensures reliability - Compact and lightweight - Easier to integrate - Low cost to fabricate - Energy sources are available widely	- Lower energy conversion efficiency - The necessity of thermal gradient - Low ZT	- Unpredictable harvested power - Uncontrolled gradient temperature	[213,223,225]
Triboelectric nanogenerator	Electrostatic induction	- Suitable for high frequency - Easier fabrication - High power density - Higher output voltage - Wide variety of materials	- Low output current - Unstable output power - High impedance - Materials - dependency	- To increase and stabilize the output power - Integration with other electronics	[226,227,228]

discusses the different techniques and its drawbacks.

5.8.4. Energy Harvesting Sources

Since mostly the WBAN nodes are battery powered, it is wise to integrate appropriate energy-harvesting modules so that wearers can go on with their daily routines without having to worry about energy depletion.

Figure 17 represents the harvesting process for a sensor node in wireless body area network process. Unlike the other WSNs, WBAN nodes can harvest energy from existing nonelectric renewable energy sources such as ambient environmental sources. Apart from them, the human body itself a very good source of energy harvesters. Based on the nature, the energy sources can be classified into different categories as shown in

Table 9. The energy sources can be classified into different categories.

Type	Energy	Source	Harvesting Techniques	Reference
Ambient Energy	Solar Energy	Sunlight, Various indoor, outdoor lighting	Photovoltaic cell	[229,230,231]
	Radio Frequency Energy	The base station, wireless networks, television towers	Rectenna
	Thermal Energy	The human body, sun, system	Thermoelectricity
	Blood Pressure Energy	Heartbeat	Piezoelectric
Biomechanical	Vibration Energy	Physical movements, walking, breathing	Piezoelectric, Electromagnetic induction	[232,233]
Biochemical	Electrochemical	Glucose	Enzymatic biofuel cell	[234,235,236,237]
		Lactate	Enzymatic biofuel cell
		Endo cochlear potential	Endoelectronics chip

:

5.8.5. Comparison of Energy Harvesters Based on Their Power Densities and Application

Table 10. Comparison of energy harvesters based on their power densities and application.

Energy Harvesters	Maximum Output Power Density (/cm2)	Application	Reference
PEG	64.9 $\mathsf{\mu }$W	Pulse sensors	[225]
TEG	9.2 mW	ECG sensors	[238]
TENG	50 mW	Glucose biosensor	[225]
Solar energy harvester	10 $\mathsf{\mu }$W (indoor)	Wearable medical sensors	[239,240,241]
	100 $\mathsf{\mu }$W (outdoor)
Biofuel cell	14 $\mathsf{\mu }$W	Self-powered biosensors	[242,243]
	64.9 $\mathsf{\mu }$W

reviewed different energy harvesters based on their maximum output power densities and application.

5.9. Challenges of WBAN

WBANs is undoubtedly a promising technology with a lot of existing issues. As time passes, challenges to the emerging technologies such as WBANs increase along with its improvement. Normally, these challenges are addressed from practicality aspects. To ensure a preferable WBAN, several ethical and technical challenges such as protocols, power issues, range etc. are yet to be solved. In this section, we have discussed some of these challenges faced by WBANs shown in

Table 11. Challenges of WBAN.

Challenges	Challenges in WBAN	References
Low powered wireless devices	Low power wireless nodes are required for the battery issue.	[244]
Efficiency	Better energy scavenging techniques are required.	[231]
Security	A high level of data security needs to be applied to intercept violations and physical attacks.	[244]
Quality of service (QoS)	The major QoS issues are limited resources, fewer capabilities, Unpredictable traffic patterns, network instability, data redundancy, node deployment	[194]
Range	Provides a small network range of a few meters	[245]
Interference	a possibility of interference as many wireless devices operates in the 2.4 GHz band [182]	[184]
No. and size of nodes	More essential and miniaturized sensors are required.	[244]
Cost	Devices used in WBANs are costly.
Awareness	Important since the body’s response changes with the context change	[244]
Networking issues	Increased network size affects the network’s routing protocols’ performance and throughput.	[194]
Mobility Support	Mobility can affect several applications.	[245]

.

6. RFID in Biomedical and Healthcare Application

Radio Frequency Identification (RFID) has become next-generation technology due to its application in different sectors [246]. Nowadays, RFID is being used for detecting and monitoring objects in various areas such as patient monitoring, vehicles, pets, assets, equipment, livestock tracking, customer service inventory, loss control, etc. Being implanted in the human body RFID also plays a vital role in the biomedical and healthcare sectors. By detecting wound (soft tissue) inside the human body and monitoring patients information as well as also tracking the assets of the hospitals it makes the health care system or treatment easier. A typical RFID system as shown in Figure 18 has three parts which include an RFID tag that is affixed to an object for identification, an RFID reader that requests data from the tag, and middleware that processes and encrypts the information that means transmits and receives radio signals to and from the readers [247].

According to a study, RFID tags are available in two types: passive and active. Passive RFID tags have no power supply and depend on fixed readers for identification and access control, whereas active RFID tags are battery-powered and are used for identification and access control. RFID bands range from high frequency to microwave. Overall, UHF RFID passive tags are the cheapest and cover a wider range of applications [246]. Another study suggests that chip-based RFID is costly whereas chipless RFID is cost-effective and does not need any maintenance. Being versatile the chipless RFID can be printed on any type of product. As chipless RFID does not need any IC chip it consumes low power than the chip-based RFID [247].

In [248], Rigelsford suggests that the passive RFID can be used to monitor patients who have experienced soft tissue trauma. Road traffic accidents, significant surgery as part of cancer treatment, or shrapnel wounds from combat or terrorist acts can all result in such injuries. The subcutaneous RFID implant passively monitors the patient’s healing rate and can be utilized to detect infection early. The passive RFID is placed closer to the wound of the soft tissue during the last stage of the surgery. After surgery when the tissue starts to get repaired then the RFID gets absorbed in the body as it is biodegradable. The rate of healing of the tissue and the degradation of the implanted device can be monitored from the outside of the human body. Thus, during the post-surgery stage, it is able to identify the infection earlier and decreases the further risk of the complications. However, the impact of the biodegradable version of the tag was not investigated [248]. Paolini et al. proposed that a wearable sensor can detect fluid such as water, ethanol, and biological fluid. At an ultra-high frequency in the presence of a special kind of fluid the sensor can detect the fluid [249].

According to a case study Filtering facepiece respirators (FFRs, often known as facemasks) which are the most popular pieces of COVID-19 personal protective equipment were sensorized using RFID-based sensors for two applications: moisture sensing and cough detection. The textile moisture sensor can distinguish between wet masks that need to be replaced and the temperature-sensing tag can track cough occurrences that are at least 1 s apart.

Wet FFRs’ filtering capacity and their usefulness in preventing infections by airborne pathogenic agents like coronaviruses may be compromised. However, understanding of this phenomenon is limited because there is no method to measure the humidity of an FFR without the use of expensive humidity sensors. As a result, creating a low-cost humidity sensor that can be integrated into the FFRs will be helpful for further investigation [250].

Panunzio proposed that utilizing the UHF RFID communication link, the SECOND SKIN project significantly aided the implementation of flexible and elastic electronics for applications on the human skin. This innovative technology can be used to monitor a range of biophysical markers on the skin interface to expose a person’s health status or enable sensory capabilities. The information gathered can be utilized for individualized diagnosis or, more recently, pandemic containment. Indeed, epidermal sensors can be accessed remotely by fixed and/or wearable readers, which should be used in conjunction with data processing and representation software. To improve robustness and precision, more testing on data processing techniques should be conducted [251].

Another study offered an RFID-based anemia detection sensor that combines Paper-based analytical devices ($\mathsf{\mu }$-PADs) with an RFID tag that is seeded with a fixed volume of a patient’s blood. Differences in red blood corpuscle composition in the blood cause controlled alterations in the RFID tag backscatter signal response. By observing these changes at the RFID reader, a healthy and anemic patient’s blood can be differentiated. This technology offers a low-cost alternative to existing anemia screening methods. However, the robustness of the device is not ensured as there is noise in the data [252].

In [253], Wang suggested using RFID technology in Hospital Information System (HIS) to realize the closed-loop procedure of a doctor’s order in order to eliminate medical errors. The entire closed-loop procedure begins with collecting the patient’s information, continues with the confirmation of the patient’s unique identification, and finally concludes with the execution of the doctor’s order. Medical personnel can directly collect and interact with the patient’s information, ensuring maximum accuracy of medical information. However, data security and privacy challenges were not introduced.

A study suggested that during the intra-operative time, to locate tiny tumors a miniature RFID tag was implanted nearby the tumors. Then a sensor antenna capable of receiving the signal from the RFID tags was utilized to determine the location of the implanted tags. For determining the distance between the sensor antenna and the target RFID tag, the strength of the signal from the target RFID tag was measured. The estimated distance can be used by the operator to locate the tag and tumors. However, the toxicology study of the materials used for the implantable RFID tags was not explored [254].

Kiani offered a closed-loop wireless power transmission system for biomedical implants. In order to save expenditure, heaviness, and area, passive RFID tags do not use batteries. So, in this case, an implant that is inductively powered is convenient to use. The factor which plays a vital role in determining the received power that can be transferred to the device is the coupling coefficient between the transmitter and receiver coils. Any changes in coil distance, alignment, and rotation cause significant changes in the amount of received power. To stabilize the device, the transmitted power is sent in such a way that the received power remains higher than its minimum level. In all conditions, the result was stable, improving the power efficiency. However, the discrete-time nature of the control loop is not considered here [255].

In [256], Baek et al. proposed a passive RFID sensor by which wireless communication is possible for E-skin application to monitor the health condition. Here a smartwatch (RFID reader) and the e-skin sensor (RFID tag) were connected by electric filed coupling mechanism through the human body. The RFID did not use the power source and used the high frequency signal. Planer coil antennas were designed with 13.57 MHz carrier signal for the sensor. Decreasing the return loss, signal loss was monitored. The high Q factor indicates a high read range which confirmed an improved efficiency. An impedance matching circuit for the coil antenna was designed to extend the read range. By increasing the read range more, the efficiency can be improved more.

Another study proposed Ormocomp PCB based body implantable RFID which is biocompatible. As traditional PCB are not biocompatible, Ormocomp PCB generated from Ormocer (a biocompatible copolymer) was used for the RFID chip. The device was made with only one biocompatible material. To make this sensor at first an ormocmp wafer is made which follows the steps of lamination of copper or aluminum metallization on one or both sides of the wafer. Then patterning of the metallization was done by the photolithography technology. Confirming the bonding of electronic components, the encapsulation was done to make the sensor biocompatible and safe for implantation This concept can be implemented in more complicated ICs and devices [257].

Yang et al. proposed an RFID where magnetic cores were used for design, fabrication, and miniaturization which can be implemented in biomedical and pharmaceutical areas. By co-designing materials and electromagnetic structure of benchmarking case of a novel flexible magnetic composite, a BaCo ferrite-silicone composite, and a UHF RFID antenna allow the miniaturization of the circuits. Here the UHF RFID antenna allowed a wider range of frequency which ensured a better output of the sensor. Different parameters such as relative permittivity and permeability of the proposed material were calculated by applying electromagnetic field which confirmed the successful implementation of magnetic composite material. This method can be used to design more complex devices [258].

Shirehjini proposed a robust system based on passive RFID to track mobile objects such as bed tracking, wheelchair tracking, patient monitoring and critical equipment tracking in an indoor environment of hospitals. RFID readers were mounted on all mobile objects of the hospital. The RFID reader components are linked to an embedded computer, which calculates position and orientation data based on the RFID reader components. Thus, it was able to maintain the track of all mobile objects of the hospital. The calculation of the system was simple and computational error was unaffected by the environmental parameters and numerical value of objects. By implementing and experimenting with different objects the system proved less error and more feasibility. However, the power performance of the system and impact of different types of floors (wooden floor, concrete floor) has not been done [259].

Zhang suggested a wireless health monitoring method based on RFID tags by which bodily oscillation especially the vibration of the hand (patient who is suffering from Perkinson’s disease) can be measured. This experiment used only commercial off-the-shelf RFID tags and measured the phase values of RF signal as well as estimated the vibration rate based on the periods in phase change. This device can measure the hand tremor of multiple persons simultaneously even when the person is not in a fixed position. The experimental results provide the accuracy rate with less error in the oscillation per second using the time-varying signal properties. However, cost analysis has not been done for the device [260].

Parlak et al. configured an RFID to ensure the quality of object use detection in a Trauma Resuscitation Bay. The device was designed with several antennas with RFID tags. To find out the best result tag placement experiment was done on different objects. High performance gain was found for attaching multiple tags. By observing different metrics, it was found that distribution distance played a vital role for Trauma Resuscitation Bay application. However, by improving the read rate the device can be updated [261].

A study suggested passive RFID based Diaper Moisture Sensors to prevent skin infection for the babies, seniors, and disabled individuals by which parents and the caregiver will be acknowledged (by instant alarm) when the diaper gets moisture. Here, the RFID reader is connected through the internet which indicates to the caretaker when the diaper gets wet. Having a wide read range, the diaper is cheap, reusable, rinseable, comfortable, and ecologically sound. However, the device can only indicate the dry and wet state, not the various level of wetness, and also experiment in the medical environment was not conducted [262].

Another study suggested an epidermal UHF RFID (placed in various parts of the human body) for monitoring body wounds, temperature measurement, and transdermal drug delivery. The tag was a loop-like epidermal tag that was inserted into the human skin and was insulated by a thin layer that supplied high gain. Depending on where (part of the body) the tag was used and the weight of the user the performance of the device varied. By using a more suitable conducting material instead of adhesive copper the efficiency of the device can be improved [263].

Brandl et al. proposed a low-cost RFID wireless sensor to monitor the duration of dental retainer usage. Retainers temperature was measured by the sensor and the data was stored in the microcontroller. As the sensor was battery dependent a power reducing network was enabled which reduced the current consumption. An ultra-high frequency was used to transfer the data from the sensor to the reader Biocompatible polymeric encapsulation was done for the device. However, the effect of the retainers depending on the age group of the patients was not monitored [264].

In [246], Haddara investigated the application and adoption of RFID in healthcare. This study suggested by adding a smart wristband with RFID tag the identification, tracking, and monitoring of patients is possible. This investigation also found the technological, financial, safety, and security constraints of deploying RFID system in the healthcare sector. Frisch deployed thousands of RFID tags into a cancer center a study investigated the effect of deployment of RFID in large-scale area. It worked on the accuracy of the data, management of assets, and impact of the multiple tags complexity [265].

7. Sub-6 GHz 5G Systems

The 5G spectrum is divided into multiple frequency bands, just like earlier generations of cellular technology. Deployment of this technology can take place throughout a vast spectrum where sub-6 GHz and mmWave are the two extremes to consider. 5G deployments require both, but each has a unique set of advantages and disadvantages.

While mmWave communication, as the name implies, is meant for substantially high frequency applications, a significant portion of devices/radio infrastructures are not well compatible for those tera-hertz level communications yet. Besides, the higher frequency communication is susceptible to more losses considering distant and wider range of coverage.

On the other hand, not being too far from the predecessors such as 4G LTE, sub-6 GHz is drawing considerable attentions as it is deemed as more practical and feasible approach towards deployment of 5G. In other words, the range of frequency of sub-6 GHz 5G is in close proximity to that of pervious technology and hence, no drastic change or adaptation will be required from the previous architecture during deployment of 5G. In addition to this, mere frequency enhancement in mmWave communication will not able to overcome the issue with range, penetration of signals through obstacles, lower coverage without significant modifications in the network’s architecture. Sub-6 GHz5G will have access to additional spectrum in the 700 MHz and 3.4–3.6 GHz frequency ranges, respectively, which indicates that this would provide a range of frequencies with higher band width and better coverage [266]. This is quite reasonable to desire to attain high speed data rate with the least possible latency with 5G mmWave technology, however, the ability of sending data quite reasonably fast to some expanded areas, including suburban and rural areas through, has made sub-6 GHz 5G more suitable and practical by most of the carriers, and service providers [267].

The frequency range between 1 and 6 GHz is known as the mid-band spectrum, and it is perfect for 5G mid-band since it can carry a significant amount of data while also transmitting over a long distance. The range 3.3–3.8 GHz, which is also called n78 band, is widely used by the majority of the countries as shown in

Table 12. The global 5G sub-6 GHz mid-band spectrum.

Countries	1–3 GHz Band	3–4 GHz Band	4–5 GHz Band
Korea	2.3–2.39 GHz	3.4–3.7 GHz, 3.7–4.0 GHz
Japan		3.6–4.1 GHz	4.5–4.9 GHz
China	2.50/2.6 GHz	3.3–3.6 GHz	4.5–5 GHz
EU		3.4–3.8 GHz
UK		3.4–3.8 GHz
Germany		3.4–3.8 GHz
France		3.46–3.8 GHz
Italy		3.6–3.8 GHz
USA	2.50/2.6 GHz	3.45–3.7 GHz, 3.7–3.98 GHz	4.49–4.99 GHz
Canada		3.475–3.65 GHz, 3.65–4.0 GHz
Australia		3.4–3.7 GHz
India		3.3–3.6 GHz

.

In this part, different types proposed antenna systems, their method of designs, their noteworthy features, performances and applications in the field of sub-6 GHz 5G will be discussed.

7.1. Single Element Antennas for Sub-6 GHz 5G Systems

Single-element antennas are considered simple and preferred due to not having issues with coupling, parasitic effect, and space, unlike in the case of multiple elements. Both frequency and pattern can be reconfigured while keeping the design simple with a single-element antenna, which is one of the most fundamental requirements of 5G communication. Numerous steps have been adopted to configure beams of the different single-element antenna in the literature. Two of the common and must-have features that a sub-6 GHz 5G antenna should have is configurability in terms of patterns and frequency [267]. Several approaches, including slots, notches, PIN diode, varactor diode, and multiple feedings techniques, have been adopted in this regard.

7.1.1. PIN Diode Switching Integration

One way of designing a reconfigurable antenna is integrating PIN diode switches (shown in Figure 19), as RF switches, both along any patch of an antenna and the stubs of the antenna [267]. The PIN diode basically varies the effective resonant length of an antenna which results frequency or pattern configuration [268].

The switches in the patch are mainly used for configuring the operating frequencies whereas, others, those are loaded with the stubs, generate a number of configured beams at some more variety of frequencies.

The other technique of PIN diode switching involves positioning four PIN diode switches on two layers of substrate separated by a 2.5 mm thick air gap and antenna is fed differentially using proximity coupling [269]. The switches are used to control the excitation of the radiating patch and feeding lines, thus, two distinct different frequency bands (2.45 GHz and 3.50 GHz) are generated. Although a more limited version of frequencies it can operate into, the peak gain could reach as much as 6.8 dBi.

Configuration through the technique of PIN diode switching could pose challenges in producing precise outcomes if insertion loss is higher than expectation and switching is not fast enough.

7.1.2. Slotted Antennas

A Number of slotted antenna designs have been proposed due to numerous advantages including wider impedance bandwidth, smaller and compact size, lower fabrication cost and, more importantly, multi-band operability [270]. Despite having higher gain at higher frequency, the antenna proposed in exhibits comparatively lower gain at the lower frequencies. This is due to the skin depth effect where current distribution is severely affected. In addition, the antenna is not well-matched with the connector impedance which results discrepancy between simulated and measured gain.

7.1.3. Multiple Feeding Techniques

This is one of the techniques of providing arbitrary beam tilting for the base station antenna which can be applied to a dual parabolic cylindrical reflector antenna system which is meant for operating at 3.3–7.0 GHz. This is performed by shifting the location of any feed away from the focus of the sub-reflector as shown in Figure 20 [271]. Changing the location of feed leads tilting the generated beam both vertically and horizontally which is done remotely electronically [271].

7.2. Overview of Recent Works on Single Element Antenna for Sub-6 G Hz 5G Applications

Table 13. Overview of proposed single antenna element systems for sub-6 GHz 5G.

Type of Antenna	Operating Band/Frequencies (GHz)	Performance Parameters	Physical Parameter (mm)	Application/s
A Printed low-profile antenna with PIN diodes loaded in the patch and stubs to provide reconfigurability [267].	2.6, 3.5, 4.2, 4.5, 5 and 5.5	Peak gain: 3.66 dBi Reconfiguring modes: 8 Efficiency (max): 78% No. of beams: 7	31 × 27	5G handheld devices
A deferentially feed antenna with two separated substrates and PIN diode switching provides reconfigurability [269].	2.45 and 3.50	Peak gain: 6.8 dBi Reconfiguring modes: 2 Efficiency (max): 69.5% No. of beams: not reported	100 × 100 × 3.2	WLAN and sub-6 GHz
A novel shaped low-profile monopole antenna with four pin diode switches able to reconfigure frequency [268].	1.8, 2.1, 2.6, 3.5, 4.8, 5, 5.6, 6.4 and 6.5	Peak gain: 3.6 dBi Efficiency: 84% Reconfiguring modes: 5 No. of beams: not reported	40 × 32 × 1.6	5G handheld devices
A multi-beam antenna consists of two parabolic reflector and multiple feeds to provide arbitrary beam tilting electronically [271].	3.3–7.0	Peak gain: 12 dBi Efficiency: not reported Reconfiguring modes: 5 No. of beams: 5	00 × 560	Base station antenna
Multi-band inverted E and U shaped antenna having two branches those hold slotted structures [270]	0.77, 1.43, 2.13, 3.48, 3.84, 5.17, and 6	Peak gain: 2.6 dBi Efficiency (max.): 87% No. of beams: Omnidirectional	30 × 30 × 1.6	Digital broadcasting, medical telemetry, UMTS, WiMAX, sub 6 GHz 5G, WLAN, fixed satellite communication
A wideband spiral shell Dielectric Resonator Antenna (SsDRA) manufactured based on bottom-up micro-Stereolithography (SLA) [273]	3.3 and 5.3	Peak gain: 4 dBi Efficiency: Not reported	Thickness: 2 mm Height: 29 mm Feed probe height: 17 mm	Not reported. (Manufacturing is more focused than applications)

summarizes a few selected articles those have developed multiple single element antenna systems for sub-6 GHz 5G applications. It is observed that the most significant drawbacks of any single element antenna system is failure to achieve optimum gain, and based on the literatures, up until now the maximum gain was found as 12 dBi [272]. In addition, channel capacity was either insignificant or unreported. Both pattern and frequency reconfigurability were achieved, however, strategies such as diode switching [268,269] and multiple feeding [271] were not quite successful in this regard. The techniques are either complex or cumbersome while offering only a limited number of beam steering options [271].

7.3. Multiple Element Antenna and Array

Enhancement in channel capacity, data rate and spectral efficiency has led researchers to switch to multiple antenna systems, such as Multiple-Input Multiple-Output (MIMO) and array antenna system. This technology is the capable of pushing the efficiency up to 80% and channel capacity can be achieved over 50/b/s/Hz [274] since multiple channels requires less power to transmit [275].

Table 14. Overview of proposed multiple antennas element systems for sub-6 GHz 5G.

Type of Antenna	No. of Elements	Operating Band/Frequencies (GHz)	Performance Parameters	Physical Parameter (mm)	Application/s
Multiband Shared-aperture slot antenna: use of single strip line and power divider for sub 6 GHz and mmWave, respectively, [279]	two	4–4.5, 3.1–3.8, 2.48–2.9, 1.82–2.14 and 1.4–1.58	Peak gain: 8.2 dBi Efficiency: Not reported Isolation: Not reported Diversity gain: Not reported Channel Capacity: Not reported ECC = 0.113	120 × 60	5G enabled devices and AP applications
4-port co-planar waveguide fed antenna with common radiator and flexible substrate [280]	Single patch-four ports	0.6–1.09, 2.6–3.4 and 4.2–7.0	Peak gain: Not reported Efficiency: 75% Isolation: >10 dB Diversity gain: Not reported Channel Capacity:29.83 bits/s/Hz ECC < 0.50	180 × 60	Emerging 0.6–1.09 GHz band, sub 6 GHz 5G near radio (NR), and Wi-Fi 6 communications
A four elements MIMO with microstrip circular patch antenna [272]	four	5.6–5.67	Peak gain: 12.4 dBi Efficiency: 85.1% Isolation: >30 dB Diversity gain: 10 (approx.) Channel Capacity: Not reported ECC: <0.005	160 × 70	5G wireless terminals
A MIMO system with coplanar-waveguide (CPW) fed antennas with polarization diversity [277]	eight	3.35–6	Peak gain: 3.5 dBi (at 6 GHz) Efficiency: 95% (at 3.4 GHz, degrades at higher frequency) Isolation: Not reported Diversity gain: Not reported Channel capacity: Not reported	75 × 150 × 1.6	5 G smartphones
A MIMO system comprised of a low profile wideband PIFA and a compact wideband monopole [274].	two	0.617–6	Peak gain: Not reported Efficiency: >68% Isolation: >26 dB Diversity gain: >9.94 dB ECC: <0.01 Channel capacity: Not reported	Not reported but distance between element is 600 mm	V2X communications (automotive)
A multiband array with T-shaped couple-fed slot elements [278]	ten	3.4–3.8 and 5.15–5.92	Peak gain: Not reported Efficiency: 82% Isolation: >11 Diversity gain: Not reported ECC: 0.05 (high band) 0.015 (low band) Channel capacity: 50–51.4 bps/Hz	16.2 × 3 × 0.8	smartphones

represents an overview of multiple antenna element system of sub-6 GHz for the implementation of 5G.

Present Shortcomings

Although as shown in Table 14, the performance of the MIMO system was found as considerably good, the number of elements used limits the overall performance. On the other hand, design complexity escalates if the number of elements is increased. One of the examples of these complexities is the high interconnection amongst the chip used in the design which is in fact, a bottleneck in the approach of designing an integrated MIMO system [276]. The current literature has addressed the following issues as summarized in Table 14.

(1)

The majority of the systems were designed with two elements only whereas increasing the number of elements has certainly been proved to be for enhanced data rate and channel capacity. There are some works on eight [277] or even ten elements [278], however either the performance has severely been degraded or some key performance parameters, including Envelope Correlation Coefficient (ECC), DG, MEG and isolation, have not been mentioned. Besides, the overall size and edge-to-edge distance between elements of the MIMO was found to be an issue while keeping the performance and compactness up to the mark simultaneously.

(2)

Consistent performance was not achieved throughout the whole band of operation as observed. For example, efficiency of the antenna was found to be degraded at the higher frequencies [279] which was the case for ECC also. The channel capacity and DG normally improves as more elements are added, however, maintaining ECC below 0.5 and better isolation between elements becomes more challenging.

(3)

Performance of the antennas, those are designed for handheld devices, is prone to degrade further. In this case, positioning of the elements of the MIMO is crucial.

(4)

Although many articles have intended to design antenna systems for a wide range of frequency in sub-6 GHz region, owing to the reasons mentioned above effective operation is not possible for at the frequencies they are designed for. In other words, the above constraints have made the MIMO antennas limited in terms of their operation.

8. Conclusions

An overview of the concepts that can be applied to wireless communication integrated with IoE is provided in this article. For the development of next-generation electromagnetic devices, transformation electromagnetic/optic theory has been presented. By implementing this theory, the meta-material devices have effectively revolutionized the network of wireless power transfer, sensors, and 5G wireless communication. Different wireless power transfer techniques have been reviewed here, which will give an overview of the recent research works. This technology can deliver power to RFID sensors and play an essential role in wireless body area networks and wireless sensor networks as an energy harvester. Hence, wireless body networks have been discussed, how the modern medical science has been changing drastically and how it is connected to the internet to establish intelligent network architecture. In the extension of the development in the healthcare system, a detailed review of recent works has been studied how RFID has made a significant advancement in the biomedical industry. Furthermore, the 5G wireless communication has revolutionized the wireless and telecommunication field. A brief study has been presented focusing on its sub-6 GHz, and different single ad multiple antenna element systems have been discussed. Finally, different algorithms have been presented to detect the signal’s direction of arrival. It can be integrated with these technologies to designate an adequately competent network to solve more tough tasks. Therefore, it can be said that these technologies are interconnected. The integration of these technologies will fulfill different aspects of wireless communication, and hence by connecting to the Internet of Everything, one can bring trajectory to this era.