1. Introduction
In recent years, a medical and healthcare Internet of Things (IoT) system has attracted attention as a means to support home medical care or a remote medical care. The system involves wearable wireless vital sign sensors or medical robots (
Figure 1) [
1,
2]. A TV-based medical service (t-health) for supportive living is proposed, which enables elderly people and physically handicapped people to easily collect basic parameter values and send them to a remote monitoring center [
1]. The authors design, implement and test the solution that provides social health services to elderly people at home based on access to smart TV technology and all services [
2]. Then, wireless body area networks (WBAN) are well-known medical and healthcare IoT systems [
3,
4,
5,
6,
7]. WBAN consist of a collection of low-power, miniaturized, invasive or non-invasive lightweight devices with wireless communication capabilities that operate near the human body. These devices are placed in, on or around the body and monitor vital information [
3,
4,
5,
6,
7]. For example, the overview of WBAN, and the recent technical and design challenges are introduced [
3,
4,
5,
6]. Comprehensive research and detailed analysis of coexistence problems and interference mitigation solutions in the WBAN are provided [
6]. The introduction of cloud support type WBAN and the main issues to be addressed for its development and management are described [
7]. By the way, IEEE 802.15.6 (Institute of Electrical and Electronics Engineers, New York, NY, USA) which is a system standard was issued in 2012 [
8]. Subsequently, system specifications for a physical layer (PHY) and a media access control layer (MAC) in smart body area networks (SmartBAN) were issued in April 2015. These specifications represent a standard for medical and other health care advanced by the European Telecommunications Standards Institute (ETSI) [
9,
10].
SmartBAN characteristics include the following. Firstly, a star-type topology is adopted, in which the hub collects data measured by the node and sends the collected data to external equipment. Secondly, complex specifications (in particular, the MAC layer), which were a problem in IEEE 802.15.6, are simplified as much as possible. Thirdly, the SmartBAN standard focuses on transmission and reception of emergency signals with low delay not considered in existing wireless communication systems. In addition, SmartBAN is standardized to satisfy the following technical requirements:
Evaluation and testing of SmartBAN have mainly focused on the MAC protocol [
11,
12,
13,
14,
15]. For example, [
12,
14] derived closed-form analytical models for the uplink transmission delay and presented an optimal inter-beacon interval frame for SmartBAN to reduce delay or energy consumption. In addition, the same authors [
12,
13,
14] evaluated downlink delay performance based on SmartBAN and improved downlink delay to adopt fixed-length exhaustive transmission [
13]. However, the SmartBAN PHY has not been sufficiently discussed.
Previously, we provided performance evaluations of an error-control scheme in the SmartBAN PHY under several conditions [
16]. In particular, we evaluated performance when Bose–Chaudhuri–Hocquenghem (BCH) codes with nearly the same redundancy as the packet repetition were applied and then compared this performance with that of the standard scheme. In addition, retransmission performance was evaluated. Numerical results indicated that retransmission substantially improved the packet error ratio and energy efficiency of the IEEE model CM3 which is a channel model of wearable WBAN [
16,
17].
In this study, first, preamble detection in the SmartBAN PHY is evaluated because other studies on the ETSI SmartBAN PHY, including our previous study, do not fully consider this point [
16,
18,
19,
20,
21,
22,
23,
24,
25]. Additionally, we propose a preamble structure to which a start frame delimiter (SFD) is added to correctly detect the header position under not only the additive white Gaussian noise (AWGN) channel (the ideal environment) but also the IEEE model CM3 (close to real environment). Several SFD candidates were selected, and their preamble detection performances are evaluated by computer simulations. The best performance is obtained when an orthogonal maximal length sequence (M-sequence) is used as the SFD under the AWGN channel and IEEE model CM 3.
As the second main contribution, we also provide a novel integrated performance evaluation of the packet error ratio (PER) and energy efficiency while taking into account preamble detection in the SmartBAN PHY. Novel numerical results by computer simulations indicate that the best performance with respect to PER is when a preamble with orthogonal M-sequences of 4 octets is used. However, for energy efficiency, better results are obtained using a preamble with orthogonal M sequences of 2 octets. Furthermore, the optimum length of the PHY packet to achieve the maximum energy efficiency with PER less than 10−2 is found by the theoretical analysis which was not conducted in previous work. Those results suggest that it is better to change the length of SFD according to channel conditions and the payload size because the maximum energy efficiency is affected by the overall packet size. They are also expected to contribute to the design of highly reliable and energy-efficient SmartBAN.
The remainder of this paper is organized as follows. In
Section 2, we summarize the SmartBAN PHY. In
Section 3, our proposed preamble structure and its performance are explained. The numerical results of an integrated performance evaluation are provided in
Section 4. Conclusions and suggestions for future research are presented in
Section 5.
5. Discussion about Optimum LPSDU
In the previous section, computer simulations were performed with the fixed un-coded . In this section, the optimum un-coded that satisfies a certain condition is discussed. Here it is assumed that the maximum energy efficiency with PER less than is achieved as a condition.
First of all,
Figure 20 shows the bit error probability (
) of the GFSK used in SmartBAN, minimum-shift keying (MSK) and 2FSK under the AWGN channel as a function of
. As show in
Figure 20, the bit error probability of the GFSK used in SmartBAN is almost the same as that of MSK. The
of MSK under the AWGN channel is expressed as follows [
34]:
Here,
is the Q function. Also, the
of MSK under the Rician fading channel is expressed as follows [
34]:
where
is the
factor (linear scale), and
is the average
. In this discussion, it is assumed that (9) is
of GFSK used in the SmartBAN PHY under the IEEE model CM3.
Then,
and
are expressed as follows [
29]:
Here
is the code-length of BCH code, and
is the number of correctable bits in the block. In the non-BCH coding case,
and
. Then, the
at a receiver can also be expressed by using
and parameters listed in
Table 2 as follows:
By using (3)–(7) and (9)–(15), the optimum
that satisfies the above condition is searched for in the full search. Here, the results of the two-octet and the four-octet orthogonal M-sequences obtained by computer simulation under the IEEE model CM3 are used as for
. Then,
is searched for multiples of 113 (no-BCH encoding case and (127,113) BCH encoding case) and 64 ((127,64) BCH encoding case) because of the information bit length of each BCH codes. In addition, the maximum
is set to
in accordance with the SmartBAN specification.
Figure 21 and
Figure 22 present the optimum
and the maximum energy efficiency under the IEEE model CM3 as a function of
. As show in
Figure 21, the optimum
of the four-octet SFD is slightly larger than that of the two-octet SFD in the cases of BCH encoding under less than 20 dB
conditions because of each
. On the other hand, the maximum energy efficiency of the two-octet SFD is larger than that of the four-octet SFD in the no BCH encoding case under over 26 dB
conditions in
Figure 22. This is because each optimum
is short, that is, it is more susceptible to the length of the SFD.
Figure 23 and
Figure 24 shows the optimum
and the maximum energy efficiency under the IEEE model CM3 as a function of the Tx-Rx distance. As shown in those figures, the optimum
and the maximum energy efficiency of the four-octet SFD are almost the same as those of the two-octet SFD in the cases of BCH encoding. On the other hand, the optimum
and the maximum energy efficiency of the four-octet SFD is much larger than those of the two-octet SFD in the no BCH encoding case. The reason is that
of the two-octet SFD is close to
, and
of no BCH encoding case is higher than that of BCH encoding case. Hence, it can be said that the two-octet SFD is better in the short optimum
case, while the four-octet SFD is better in the long optimum
case from
Figure 21,
Figure 22,
Figure 23 and
Figure 24.
6. Conclusions
In this manuscript, we evaluated preamble detection in the ETSI SmartBAN PHY and proposed a modified preamble structure. Specifically, an SFD was added between the two-octet preamble and the PLCP header. The proposed preamble structure is compatible with the SmartBAN standard. This is because the general framework of the packet structure is not changed, and the only minor modification is made. Computer simulations indicated that the preamble with an SFD consisting of the four-octet orthogonal M-sequence has better detection performance than SmartBAN and similar approaches, in particular, under poor channel conditions with IEEE model CM3. In addition, integrated performance evaluation with respect to PER and energy efficiency considering preamble detection in SmartBAN PHY was conducted. According to computer simulation results, the case with an SFD whose 4 octets and with an orthogonal M-sequence exhibited better PER performance, while larger energy efficiency was achieved in the case of an SFD whose 2 octets and with an orthogonal M-sequence. Furthermore, it was possible to find the optimum and the SFD satisfying the maximum energy efficiency with PER less than from theoretical formulas. Those results suggested that it is better to change the length of SFD according to channel conditions and the optimum . For example, the two-octet SFD was better in case that the channel was high and the optimum was short. On the other hand, the four-octet SFD was better in the long optimum case.
In future research, other error-control schemes and access protocols will be evaluated and analyzed. In addition, the proposed system will be implemented in a real-world application. As the future prospects, we are aiming to amend the SmartBAN standard based on our proposed system. For example, it is conceivable to amend so that it can be selected whether or not to use the SFD, and the length of the SFD can be selected according to the node priority.