Immersive communications is a key use case for 6G where applications require reliable latency-bound media traffic at a certain data rate to deliver an acceptable User Experience (UX) or Quality-of-Experience (QoE). The Quality-of-Service (QoS) framework of current cellular systems (4G and 5G) and prevalent network congestion control algorithms for latency-bound traffic like L4S typically target network-related Key Performance Indicators (KPIs) such as data rates and latencies. Network capacity is based on the number of users that attain these KPIs. However, the UX of an immersive application for a given data rate and latency is not the same across users, since it depends on other factors such as the complexity of the media being transmitted and the encoder format. This implies that guarantees on network KPIs do not necessarily translate to guarantees on the UX. In this paper, we propose a framework in which the communication network can provide guarantees on the UX. The framework requires application servers to share real-time information on UX dependency on data rate to the network, which in turn, uses this information to maximize a UX-based network utility function. Our framework is motivated by the recent industry trends of increasing application awareness at the network, and pushing application servers towards the edge, allowing for tighter coordination between the servers and the 6G system. Our simulation results show that the proposed framework substantially improves the UX capacity of the network, which is the number of users above a certain UX threshold, compared to conventional rate control algorithms.

In this paper, we are interested in high-quality imaging of still objects with only received power measurements of off-the-shelf WiFi transceivers. We show that the scattered WiFi signals off of objects carry much richer information about the edges of the objects than the surface points. Based on this observation, we then propose a completely different way of thinking about this imaging problem. More specifically, we propose Wiffract, a new foundation for imaging objects via edge tracing. Our approach uses the Geometrical Theory of Diffraction (GTD) and the corresponding Keller cones to image edges of the object. We extensively validate our approach with 37 experiments in three different areas, including through-wall scenarios. We take developing a WiFi Reader as one example application to showcase the capabilities of our proposed pipeline. More specifically, we show how our approach can successfully image several alphabet-shaped objects. We further show that our approach enables WiFi to read, i.e., correctly classify the letters, with an accuracy of 86.7%. Finally, we show how our approach enables WiFi to image and read through walls, by imaging the details and further reading the letters of the word "BELIEVE" through walls. Overall, our proposed approach can open up new directions for RF imaging.

The detection of nocturnal seizures in epilepsy patients is essential, both for the quick management of the seizure complications, and for the assessment of the ongoing seizure treatment. Traditional seizure detection products (e.g., wearables), however, are either very costly, uncomfortable, or unreliable. In this article, we then propose to utilize everyday WiFi signals for robust, fast, and noninvasive detection of nocturnal seizures. We first present a new and rigorous mathematical characterization for the spectral content/bandwidth of the WiFi signal, measured on a WiFi device placed near a sleeping patient, during different kinds of sleep motions: seizures, normal movements (e.g., posture adjustments), and breathing. Based on this mathematical modeling, we propose a novel pipeline for processing the received WiFi signals to robustly detect all nocturnal nonbreathing movements, and then classify them into normal body movements or seizures. In order to validate this, we carry out extensive experiments in seven different typical bedroom locations, where a set of 20 actors simulate the state of having seizures (a total of 260 instances), as well as normal sleep movements (a total of 410 instances). Our proposed system detects 93.85% of the seizures with a mean response time (MRT) of only 5.69 s since the onset of the seizure. Moreover, our proposed system achieves a probability of false alarm of only 0.0097, when classifying normal sleep movements. Overall, our new mathematical modeling and experimental results show the great potential the ubiquitous WiFi signals have for detecting nocturnal seizures, which can provide better support for epilepsy patients and their caregivers.

In this paper, we are interested in the problem of counting a crowd of stationary people (i.e., seated) using a pair of WiFi transceivers. While the people in the crowd are stationary, i.e. with no major body motion except breathing, people do not stay still for a long period of time and frequently engage in small in-place body motions called fidgets (e.g., adjusting their seating position, crossing their legs, checking their phones, etc). In this paper, we propose that the aggregate natural fidgeting and in-place motions of a stationary crowd carry crucial information on the crowd count. We then mathematically characterize the Probability Distribution Function (PDF) of the crowd fidgeting and silent periods (which we can extract from the received WiFi signal) and show their dependency on the total number of people in the area. In developing our mathematical models, we show how our problem of interest resembles a several-decade-old M/G/∞ queuing theory problem, which allows us to borrow mathematical tools from the literature on M/G/∞ queues. We extensively validate our proposed approach with a total of 47 experiments in four different environments (including through-wall settings), in which up to and including N = 10 people are seated. We further test our system in different scenarios, and with different activities, representing various engagement levels of the crowd, such as attending a lecture, watching a movie, and reading. Moreover, we test our proposed system with different number of people seated in several different configurations. Our evaluation results show that our proposed approach achieves a very high counting accuracy, with the estimated number of people being only 0 or 1 off from the true number 96.3% of the time in non-through-wall settings, and 90% of the time in through-wall settings. Our results show the potential of our proposed framework for crowd counting in real-world scenarios.

In this article, we are interested in through-wall gait-based identification of multiple people who are simultaneously walking in an area, using only the WiFi magnitude measurements of a small number of transceivers. This is a considerably challenging problem as the gait signatures of the walking people are mixed up in the WiFi measurements. In order to solve this problem, we propose a novel multidimensional framework, spanning time, frequency, and space domains, that can separate the signal reflected from each walking person and extract its corresponding gait content, in order to identify multiple people through walls. To the best of our knowledge, this is the first time that WiFi signals can identify multiple people in an area. We extensively validate our proposed system with 92 test experiments conducted in four different areas, where the WiFi transceivers are placed behind walls, and where two or three people (randomly selected from a pool of six test subjects) are walking in the area. Our system achieves an overall average accuracy of 82% in correctly identifying whether a person walking in the test experiment (referred to as a query) is the same as a candidate person, based on 6404 query-candidate test pairs. It is noteworthy that none of the test subjects/areas has been seen in the training phase.

In this paper, we are interested in passively tracking multiple people walking in an area, using only the magnitude of WiFi signals from one WiFi transmitter and a small number of receivers (configured as an array) located on one side of the area. Past works on RF-based tracking either track only a single moving person, use a large number of transceivers surrounding the area to track multiple people, or use additional resources like ultra-wideband signals. Furthermore, magnitude-based tracking provides an attractive feature that additional receiver antennas can easily be added to the antenna array as needed, without the need for phase synchronization, since the magnitude can be measured independently on the different antennas. In this paper, we then propose a new framework that uses only the magnitude of WiFi signals and expresses it in terms of the angles of arrival of signal paths at the receivers as well as the motion parameters of the virtual arrays emulated by the moving people. We then use a two-dimensional MUltiple SIgnal Classification (MUSIC) algorithm to estimate the aforementioned parameters, and further utilize a Particle Filter with a Joint Probabilistic Data Association Filter to track multiple people walking in the area. We extensively validate our proposed framework in both indoor and outdoor areas, through 40 experiments of tracking 1 to 3 people, using only one transmit antenna and three laptops as receivers (a total of four off-the-shelf Intel 5300 WiFi Network Interface Cards (NICs)). Our results show highly accurate tracking (mean error of 38 cm in outdoor areas/closed parking lots, and 55 cm in indoor areas) using minimal WiFi resources on only one side of the area.

In this paper, we propose XModal-ID, a novel WiFi-video cross-modal gait-based person identification system. Given the WiFi signal measured when an unknown person walks in an unknown area and a video footage of a walking person in another area, XModal-ID can determine whether it is the same person in both cases or not. XModal-ID only uses the Channel State Information (CSI) magnitude measurements of a pair of off-the-shelf WiFi transceivers. It does not need any prior wireless or video measurement of the person to be identified. Similarly, it does not need any knowledge of the operation area or person's track. Finally, it can identify people through walls. XModal-ID utilizes the video footage to simulate the WiFi signal that would be generated if the person in the video walked near a pair of WiFi transceivers. It then uses a new processing approach to robustly extract key gait features from both the real WiFi signal and the video-based simulated one, and compares them to determine if the person in the WiFi area is the same person in the video. We extensively evaluate XModal-ID by building a large test set with $8$ subjects, $2$ video areas, and $5$ WiFi areas, including 3 through-wall areas as well as complex walking paths, all of which are not seen during the training phase. Overall, we have a total of 2,256 WiFi-video test pairs. XModal-ID then achieves an $85%$ accuracy in predicting whether a pair of WiFi and video samples belong to the same person or not. Furthermore, in a ranking scenario where XModal-ID compares a WiFi sample to $8$ candidate video samples, it obtains top-1, top-2, and top-3 accuracies of $75%$, $90%$, and $97%$. These results show that XModal-ID can robustly identify new people walking in new environments, in various practical scenarios.

In this paper, we are interested in estimating the angle of arrival (AoA) of all the signal paths arriving at a receiver array using only the corresponding received signal magnitude measurements (or, equivalently, the received power measurements). Typical AoA estimation techniques require phase information, which is not available in some WiFi/Bluetooth receivers, and is further challenging to properly measure in a synthetic antenna array due to synchronization issues. In this paper, we then show that AoA estimation is possible with only the received signal magnitude measurements. More specifically, we first propose a framework, based on the spatial correlation of the received signal magnitude, to estimate the AoA of signal paths from fixed signal sources (both active transmitters and passive objects). Next, we extend our AoA estimation framework to a dual setting, and further utilize a particle filter, to show how a moving target (both active transmitters and passive robots/humans) can be tracked, based on only the received signal magnitude measurements of a small number of fixed receivers. We extensively validate our proposed framework with several experiments (total of 22), in both closed and open areas. More specifically, we first utilize a robot to emulate an antenna array, and estimate the AoA of active transmitters, as well as passive objects using only the received WiFi signal magnitude measurements. We next validate our tracking framework by using only three off-the-shelf WiFi devices as receivers, to track an active transmitter, a passive robot that writes the letters of IPSN on its path, and a walking human. Overall, our results show that AoA can be estimated, with a high accuracy, with only the received signal magnitude measurements, and can be utilized for high quality angular localization and tracking.