Medical imaging is one of the most common techniques for biomedical diagnosis to identify tumor or abnormality in the human body. Recently, the excessive use of some imaging modalities are raising issues regarding exposure to radiation and healthcare costs [1
]. Radiation from conventional medical imaging system has become a critical health concern due to excessive radiation exposure [2
]. Overuse of medical imaging-based diagnosis has brought increased exposure to ionizing radiation along with the increased danger of unwanted effects on human body [3
]. Significant number of cancer deaths and cancer caused by medical imaging process that uses ionizing radiation, reported every year in the U.S. [4
]. Researchers are trying to find alternatives to conventional diagnosis systems such as magnetic resonance imaging, computed tomography scan and other methods because they suffer from false-positive accuracy rate. Moreover, these processes are time-consuming and cannot facilitate portability [5
] and most importantly, health concerns due to ionized exposure. For decades, these medical imaging systems have been used to detect any abnormality in the human body. Recently, the adverse impact on the human body and reliability issues raised by conventional imaging tools are being taken seriously [5
Healthcare diagnosis tools based on microwaves are being given priority by researchers. Microwave signal contrast between the electrical characteristics of human tissues can easily be distinguished by microwave antenna sensors. Radiated and Scattered power are received by one or more antenna sensors in microwave imaging. Microwave-based portable medical diagnosis tools have the potential to save lives by utilizing microwave sensor antennas that perform well.
The concept of abnormality detection in the human body using microwave imaging sensors is based on the variation of contrast in dielectric properties in malignant and healthy tissue [6
]. The water content of every biological tissue has different electrical properties. Change in dielectric properties can easily be sensed by microwave sensing antennas. Researchers specified some antenna properties as requirements for microwave imaging—unidirectional radiation pattern, a frequency range within 1–4 GHz and small dimensions to make the system portable are the most desired specifications [5
Researchers are currently working to achieve all the antenna requirements for microwave imaging. Different types of microwave antennas have been reported for microwave-based biomedical diagnosis. They are mostly reported for human-head imaging and breast-cancer imaging [7
]. In References [12
], two antennas are presented for a heart failure detection system. In Reference [14
], a folded and meandered loop antenna with three-dimensional structure was presented for medical diagnosis. Moreover, it was also proposed for the application of radio-frequency identification, ultrahigh frequency (UHF) TV channels and others. The antenna in Reference [10
] achieved good operating bandwidth where the presented antenna in Reference [5
] has low gain. Although, when antenna operates in lower microwave frequency good signal penetration can be achieved [14
] but operating frequency in less than 1 GHz might blur the result of imaging [16
]. A unidirectional antenna for microwave application was presented in Reference [17
], that operates at 1.70–1.77 GHz with a pick gain of 3.42 dB. In Reference [7
] the presented antenna covers good operating band. However, it is low in gain and the radiation pattern is not unidirectional which are key requirements. Moreover, the design is expensive substrate based where expenses in microwave imaging system in biomedical diagnosis is an issue as other costly imaging modalities are out of reach of large masses in developing countries [18
]. A lens-loaded Vivaldi antenna has been introduced in Reference [19
], which operates in ultra-wide band but it is large in dimension, gain is low at a lower microwave frequency with omnidirectional radiation. A unidirectional Vivaldi antenna has been introduced in Reference [20
]. Though these antenna can cover a good operating band, the height of the Vivaldi antenna are usually high enough that it could be an issue while placement of imaging systems and substrate is also expensive. An interesting design of a dipole antenna was presented in Reference [21
] for microwave imaging. However, the antenna is too high, the design is complex and the radiation pattern is not unidirectional, which is highly desired for microwave imaging. Though the peak gain of the antenna is 6.2 dB at 4 GHz but the gain is less than 3 dB at lower frequency. Another important factor in terms of better signal penetration in microwave imaging is high directivity. An antenna’s unidirectional characteristics can be validated by a high Front-to-back ratio, where the presented antenna in Reference [7
] obtained a front to back ratio value of 9.
In this paper, a modified stacked type PIFA antenna as a microwave imaging sensor with a folded radiating structure for biomedical diagnosis is introduced. The significance of the proposed antenna are that it uses a modified stacked shape that can use full volume for impedance matching, providing a good effective electrical length for achieving resonance at a lower microwave frequency with high gain and the design facilitates an easier fabrication process for mass production (Stamping metal sheet method/3D printing) with an extremely low cost because the used copper material is very inexpensive. The proposed antenna achieved a maximum realized gain of 4.5-dB with a front-to-back ratio value of 15. Initially, the proposed antenna sensor was designed and analyzed in Computer Simulation Technology (CST) Microwave Studio. Then, it is manually fabricated using inexpensive copper sheet (0.2-mm-thick). The reflection coefficient performance of the fabricated prototype was tested using a PNA network analyzer. Finally, a computational sensitivity test of the proposed antenna as microwave imaging sensor has been carried out. Initially, the sensitivity test was performed using a computational tissue phantom based on a reflection and transmission coefficient. After that, a computation microwave imaging scenario has been performed to detect tumors in a three-layered rectangular human tissue phantom using the proposed microwave sensor antenna.
2. Design and Methodology
The proposed modified PIFA antenna has certain advantages over two-dimensional antennas.
The design can efficiently utilize full volume by stacking technique that can provide good effective length. Moreover, the fabrication process can be easier for mass production, using the stamping metal-sheet or three-dimensional (3D) printing technology. Figure 1
shows the design configurations of the antenna. The antenna configuration has been presented in different view for better understanding. The value of the proposed design parameters of the antenna are L = 70 mm, W = 60 mm, a = 20 mm, b = 40, c = 36 mm, d = 5 mm, h1 = 10 mm and h = 27.5 mm.
The aforementioned parameters are distinctively studied keeping the others fixed to investigate how the dimensional parameters influence the operating frequency of the proposed antenna. The folding technique is implemented for size reduction and getting a maximum effective electrical length. Moreover, the occurrence of phase alternation due to the folding enhances the directivity. We can see that the main radiating structure of the antenna can be realized using the folding technique, as shown in Figure 1
c. To elevate the unidirectional property of the antenna, two reflector walls at both sides are used. To decrease the effect of spurious radiation that occurs in microstrip feeding, coaxial feeding is applied to the antenna.
A prototype of the antenna was fabricated using the parameters listed in Table 1
. The fabricated antenna prototype is shown in Figure 2
. The main radiating element of the antenna is taken from a 0.2 mm thick copper plate and resized accordingly. Lastly, it has been folded to provide the shape illustrated in the design. The reflector walls standing to the left and right side of the main radiator are also shaped from copper sheets and they are soldered on the ground plane.
3. Results and Discussion
To validate the operating principle of the simulated design of the proposed modified PIFA antenna has been fabricated and measured. Measurement of the reflection coefficient was performed using the PNA network analyzer. Figure 3
depicts the reflection coefficient the proposed antenna. The antenna achieved its operating band at 1.55–1.68 GHz, with 130-MHz bandwidth. In the simulation, the antenna exhibited resonance at 1.61 GHz. The resonant frequency right-shifted from 1.61 to 1.63 GHz in the measurement. The operating bandwidth remained almost the same in the measurement scenario. Overall, the simulation and measurement results of the reflection coefficient are in good agreement. A slight mismatch occurred due to manual fabrication.
The surface-current distribution of the antenna is shown in Figure 4
. The operating band of the antenna is elevated by the folded structure because of the coupling between the upper radiating arms lower radiating element. A strong current can be noticed at the from lower to upper radiating element of the modified PIFA radiating structure. It plays a vital role in achieving the unidirectional radiation property. The dominant current travels from the lower radiating element to the upper radiating arms, which increases the effective radiating length and allows the antenna to resonate at a lower microwave frequency band.
The far-field characteristics of the proposed antenna were achieved using the Satimo near-field measurement system, as shown in Figure 5
. The far-field radiation characteristics of the antenna at phi = 0° and phi = 90° are shown in Figure 6
, which clearly indicate that the antenna has a directional radiation pattern. The main lobe directivity magnitude is 4.58 dBi. The antenna obtains a 3-dB angular beam width of 130°. The results of the front-to-back ratio and surface-current distribution validate the radiation patterns shown in Figure 6
. Cross polarization is negligible in the phi = 0 plane. It can be seen that there is good agreement between the measured and simulated results. Polarization is satisfactory with stable radiation pattern over the bandwidth. The radiation pattern of the antenna remains unidirectional over the whole operating band, which can be validated from the measured 3D radiation patterns shown in Figure 7
The current distribution remains almost the same throughout the whole operating frequency, as shown in Figure 4
, which in turn maintains the existence of the directive radiation characteristics. The measured gain and front-to-back ratio are shown in Figure 8
a. The antenna achieves a measured realized gain of 4.5 dB and the front-to-back ratio of 15. Moreover, the front-to-back ratio value validates the directivity radiation characteristics of the antenna. The fabricated antenna prototype achieves approximately 72% efficiency.
The microwave imaging simulation was performed using the proposed antenna as an imaging sensor. Figure 9
shows the simulation configuration of a typical microwave-imaging system to detect an abnormality in a rectangular human-tissue phantom (Figure 9
a). The phantom consists of skin, fat and muscle. The dielectric constant of skin, fat and muscle are 37, 5.3 and 52 respectively. The phantom is placed between two antennas. The distance from the sensor backplane to the phantom model is 40 mm. A cyst is placed inside the tissue model, which acts as stroke or blood clot and a dielectric constant of 67 is considered and diameter is 10 mm. In this simulation, one antenna is the transmitter and the other one acts as the receiver. This study is performed to validate the performance of the proposed antenna as microwave imaging sensor. The data has been obtained using the raster scanning method [22
]. Normalized scattered parameter data have been used to construct the image in Figure 10
a. The existence of the tumor can easily be identified by the reddest part of the image. A transmission and reflection coefficient plot are presented in Figure 10
b. The variation of S-parameters clearly depicts the sensitivity of the antenna.
Another performance analysis has also been presented in Figure 11
. Three tumors of different sizes and dielectric permittivity are placed at different positions within the human tissue. Tumor 1, tumor 2 and tumor 3 have dielectric constants of 50, 40 and 70, respectively. From Figure 11
c it can be seen that three different sized tumors with different value of dielectric constant has been identified in the normalized image. Tumor 1 and 3 are clearly identified. But, tumor 2 is a little blurry because of its tiny size of 5 mm, as the step width of the scan was 2.5 mm.
The specific absorption rate (SAR) analysis was performed to observe the electromagnetic wave absorption on the phantom, as shown in Figure 12
a, where the simulated power of 0.5 W was considered. It can be seen that the maximum 10 g SAR value of 1.68 W/Kg was experienced, which complies with the IEEE and ICNIRP guidelines. The power density over the human tissue phantom was analyzed, as shown in Figure 12
b. We can see that a large amount of power is absorbed on the tumor concealed area due to the high dielectric property of the tumor tissue. In Figure 11
b, the round and most red area shows the presence of tumor and the power density inside the phantom appears more similar to the shape of the considered tumor in the simulation.
Finally, a microwave imaging experiment was performed using the proposed sensor antenna. Two realistic breast phantoms have been used in the experiment. One of the breast phantom contains a tumor of dielectric permittivity 67. The other phantom is free from abnormality. The experiment was conducted to verify the antenna’s potential to detect the abnormality in the tumor contained breast phantom. Table 1
illustrates the important specifications of the breast phantom.
The experimental setup has been presented in Figure 13
. The experiment setup consists of two sensor antennae, the breast phantom, a motor control unit, vector network analyzer and a computer. The computer is interfaced with a vector network analyzer to collect scattered parameters of the antennas. Two antenna sensors are connected with two different ports of the VNA. The breast phantom is placed on a tray that can be rotated by a stepper motor, placed in between two sensor antennas. As the breast phantom rotates, the computer collects the scattered parameters from the vector network analyzer. For further understanding, Figure 13
c depicts the block diagram of the experimental setup. After post processing the data, the imaging results of the breast phantoms are presented in Figure 14
In Figure 14
, the internal structure of the phantom breast is depicted where scattering of the electromagnetic wave is highlighted. Figure 14
b clearly confirms the presence of abnormality when compared with Figure 14
a. Figure 14
a,b represents the experimental results of breast phantom without tumor and with tumor respectively. In Figure 14
a, the electromagnetic signal left no significant variation but rather passed right through the healthy breast phantom. But, significant variations can be observed in Figure 14
b, which assures the existence of a tumor or abnormality in the second breast phantom.
The proposed antenna is compared with other antennas in Table 2
. Comparing the proposed antenna with others, it is observed that this antenna is an eligible candidate as a microwave imaging sensor in terms of compact size, lower frequency operation, high gain and unidirectional radiation, which are highly desired specifications.