We present Tesla-Rapture, a gesture recognition system for sparse point clouds generated by mmWave Radars. State of the art gesture recognition models are either too resource consuming or not sufficiently accurate for the integration into real-life scenarios using wearable or constrained equipment such as IoT devices (e.g., Raspberry PI), XR hardware (e.g., HoloLens), or smart-phones. To tackle this issue, we have developed Tesla, a Message Passing Neural Network (MPNN) graph convolution approach for mmWave radar point clouds. The model outperforms the state of the art on three datasets in terms of accuracy while reducing the computational complexity and, hence, the execution time. In particular, the approach, is able to predict a gesture almost 8 times faster than the most accurate competitor. Our performance evaluation in different scenarios (environments, angles, distances) shows that Tesla generalizes well and improves the accuracy up to 20% in challenging scenarios, such as a through-wall setting and sensing at extreme angles. Utilizing Tesla, we develop Tesla-Rapture, a real-time implementation using a mmWave Radar on a Raspberry PI 4 and evaluate its accuracy and time-complexity. We also publish the source code, the trained models, and the implementation of the model for embedded devices.

Recognition of the context of humans plays an important role in pervasive applications such as intrusion detection, human density estimation for heating, ventilation and air-conditioning in smart buildings, as well as safety guarantee for workers during human-robot interaction. Radio vision is able to provide these sensing capabilities with low privacy intrusion. A common challenge though, for current radio sensing solutions is to distinguish simultaneous movement from multiple subjects. We present an approach that exploits antenna installations, for instance, found in upcoming 5G technology, to detect and extract activities from spatially scattered human targets in an ad-hoc manner in arbitrary environments and without prior training of the multi-subject detection. We perform receiver-side beamforming and beam-sweeping over different azimuth angles to detect human presence in those regions separately. We characterize the resultant fluctuations in the spatial streams due to human influence using a case study and make the traces publicly available. We demonstrate the potential of this approach through two applications: 1) By feeding the similarities of the resulting spatial streams into a clustering algorithm, we count the humans in a given area without prior training. (up to 6 people in a 22.4 m2area with an accuracy that significantly exceeds the related work). 2) We demonstrate that simultaneously conducted activities and gestures can be extracted from the spatial streams through blind source separation.

We propose Chorus, a new ultra-wideband (UWB) localization scheme in which the target device computes the time difference of arrival (TDOA) of signals sent concurrently by localization anchors in known positions. This scheme, similar to GPS, is the opposite of existing TDOA schemes for UWB, where the target transmits the signal and anchors compute the time difference. This reversed perspective enables several advantages in Chorus, including support for countless targets.

In the 2.4 GHz unlicensed spectrum, the coexistence of WiFi, Bluetooth and IEEE 802.15.4 devices generates increased channel contention. Notably, low-power wireless networks experience packet loss and delays due to interference. To improve the performance of low-power wireless networks under interference, we propose a data driven proactive approach based on interference modeling for white space prediction. We leverage statistical analysis of real-world traces from two indoor environments characterized by varying channel conditions to identify interference patterns. We characterize interference in terms of Inter-Arrival Time (IAT) and number of interfering signals and use a Gaussian Mixture Model (GMM) to accurately estimate the interference distribution as observed by the low-power wireless nodes. Then, we use a Hidden Markov Model (HMM) for white space prediction. Our validation w.r.t. real-world traces from two environments show that our GMM model can estimate interference with an accuracy higher than 94:7%. Moreover, the white space prediction evaluation shows an average accuracy of 97:7% and 89:5% across the two environments.

Cross Technology Interference (CTI) is a prevalent phenomenon in the 2.4 GHz unlicensed spectrum causing packet losses and increased channel contention. In particular, WiFi interference is a severe problem for low-power wireless networks causing a significant degradation of the overall performance. We propose here a proactive approach based on WiFi interference modeling for accurately predicting transmission opportunities for low-power wireless networks. We leverage statistical analysis of real-world WiFi traces to learn aggregated traffic characteristics in terms of Inter-Arrival Time (IAT) that, once captured into a specific 2nd order Markov Modulated Poisson Process (MMPP(2)) model, enable accurate estimation of interference. We further use a hidden Markov model (HMM) for channeloccupancy prediction. We evaluated the performance of: i) the MMPP(2) traffic model w. r. t. real-world traces and an existing Pareto model for accurately characterizing the WiFi traffic and, ii) compared the HMM based white space prediction to random channel access. We report encouraging results for using interference modeling for white space prediction.