Petar Popovski, Editor in Chief of IEEE JSAC
Published: 19 Aug 2022
The July 2022 issue of IEEE JSAC, same as the previous one, deals with Integrated Sensing and Communication (ISAC). In order to diversify with respect to the trend of novel paradigms for communication and sensing pursued in 6G, in this edition of the blog we will feature a practical method for sensing that is utilizing the radio waves of an existing Wi-Fi system.
We usually treat GPS localization as a service that is constantly available; however, GPS signals are vulnerable to blockages and may not offer the desired precision in e.g. indoor scenarios. To address this problem, a number of research works and engineering solutions rely on localization based on WiFi signals. This is convenient, considering the widespread deployment of WiFi access points as well as many ways to extract information from the wireless WiFi signals through received signal strength, time, angle, or frequency. The authors of WheelLoc add a very creative twist to WiFi-based localization by leveraging the regularity of wheel rotation in bikes or vehicles. As Figure 1 shows, antennas are mounted on the wheel and, when these antennas rotate, they exhibit a behavior that can be represented by a large virtual antenna array, whose Angle of Arrival (AoA) and Time of Arrival (ToA) are utilized to make the localization accurate. Unlike the previous works that have controlled the rotation to aid localization, here the approach is not to interfere with the rotation of the wheel, that is the behavior of the biker or a driver, but to get the rotation as an uncontrollable input parameter.
The hardware for WheelLoc is built by using on commercial-of-the-shelf (COTS) components. A key ingredient is the triangular antenna structure (see Figure 1) that rotates with the wheel and extracts wireless channel information from the received WiFi signals. The system also comprises inertial sensors that collect motion information about the antennas. The localization algorithm consists of four subcomponents: (1) StatLoc: used to do localization when the wheel is static; (2) MoveLoc: Trace-based Synthetic Aperture Radar (TSAR), suitable for flat terrain; (3) DiffLoc: Circular Synthetic Aperture Radar (CSAR); and (4) InerCalib: Compensation for Non-Line-of-Sight (NLoS) signals using the inertial sensors. Overall, the method achieves decimeter-level localization accuracy.
JSAC: In your approach you have used Wi-Fi for sensing and there are also a number of other sensing applications that rely on Wi-Fi. However, the research towards 6G looks into new waveforms for sensing and communication. Which are the disadvantages of Wi-Fi-based sensing that can be overcome by these new waveforms?
Due to the wide deployment of Wi-Fi, it has been applied in various wireless sensing scenarios. Our approach utilizes the original waveform of Wi-Fi and the natural moving properties (rotations and translations) of wheels. It makes our design realize the ISAC (integrated sensing and communication) in a quite low-cost manner without complicated modifications to communication systems or devices, which, of course, is one of the main reasons we still choose Wi-Fi as our sensing method.
We have learnt that researches have been conducted on new waveforms in 6G for sensing and communication. For example, the Orthogonal Time Frequency Space (OTFS) has been proposed as a new waveform in 6G and a natural candidate for ISAC. It multiplexes the information symbols in the delay-Doppler (DD) domain, which can provide quasi-static, compact, and separable features that can overcome the disadvantages of Wi-Fi-based sensing. To be specific, the channel response in the DD domain is relatively static, so the OTFS-based sensing can support highly dynamic scenarios when targets move at a high speed. And only a pair of moving speed and latency information can describe the channel. Additionally, due to different time variances, multiple paths are separable in the DD domain compared with the TF (time-frequency) domain. It can solve the multi-path impact in wireless sensing and has the potential to increase sensing accuracy and granularity.
Although the 6G new waveform has its own advantages for sensing and communication, it is provided and managed by operators. The 6G-based ISAC services thus are decided by operators rather than clients. By contrast, Wi-Fi works in the unlicensed spectrum, and users can customize the deployment on their LANs. We also noticed that Wi-Fi is also moving towards Wi-Fi7, which offers up to 4096-QAM modulation and 320MHz bandwidth, still reflecting a bright future of Wi-Fi-based sensing.
JSAC: What do you think is the weak point of your approach? Was there a criticism from the reviewers that led you to rethink part of your work?
The weak point of our approach is the dependence on the prior knowledge of AP (access point) locations. In our preliminary experiments, we evaluated our designs in an outdoor region with 4 AP deployments, so the location measurements of these APs did not take much time or energy. But it will cause trouble when the system is deployed on a larger scale, such as on the whole campus where there are more than 1000 APs.
The criticism from the reviewers can indeed help in further improving our work. We found that most reviewers are concerned about the location accuracy achieved by our design since it mainly relies on the phase information of the received signals. However, the phase information can be easily corrupted by multi-path or NLoS (non-line-of-sight) interference, especially in a dense area with multiple persons or vehicles. In our paper, we have designed the InerCalib module to deal with these potential factors, and we are also working on new techniques to deal with these problems when the system is deployed to a larger space.
Statements and opinions given in a work published by the IEEE or the IEEE Communications Society are the expressions of the author(s). Responsibility for the content of published articles rests upon the authors(s), not IEEE nor the IEEE Communications Society.