A novel hybrid machine-learning approach is proposed in this work to carry out robust multi-people tracking using a millimeter-wave (mmWave) radar in complex indoor environments. The proposed new system including the adaptive hybrid clustering technique leverages both density-based spatial clustering (DBSCAN) and expectation-maximization (EM) schemes over the received radar point-clouds to separate individual human trajectories reliably. Meanwhile, the Hungarian algorithm incorporation with the Kalman filter is employed for tracking of individual persons. Furthermore, a condition-number-based outlier detector is designed to filter error-prone radar data to improve the tracking accuracy. Realworld experiments are conducted in various indoor scenarios and the results demonstrate that our proposed new system can achieve an average Euclidean-distance error (EDE) of 38.19cm in the presence of a single person and that of 44.70cm in the presence of two persons in our laboratory with the area of 300cm×300cm. Our proposed new multi-subject tracking approach also outperforms the existing methods in terms of EDE.

As unmanned aerial vehicles (UAVs) are employed for many applications, UAV identification becomes critical to air traffic control nowadays. Conventional radio-frequency (RF) based UAV identification schemes can correctly recognize a UAV according to the acquired sufficient training RF signal data from a pre-specified training candidate set. However, they may often fail when the model of a test UAV is not included in such a training candidate set and/or the training data are quite limited. To address the aforementioned practical challenges, a novel RF-based open-set few-shot UAV recognition technique is introduced in this work. In our proposed new approach, an RF signal of interest, such as a UAV control signal and a digital data-transmission (DDT) signal, is first sensed and segmented in the UAV frequency band using the corresponding short-time energy-to-spectral entropy ratio (ST-ESER). Then, the cyclic spectrum analysis is applied to the selected RF signal segments to construct the corresponding polyspectra, which are further utilized to establish the cyclic-paw-print (CPP) tensors. Moreover, we design a novel deep-learning (DL) network, namely transformer-attention squeeze-excitation network (TASE-Net), by fusing the transformer-enhanced squeeze-and-excitation (SE) model and the Gaussian mixture model (GMM) into the residual network. The TASE-Net can excel in global feature modeling and unknown class detection simultaneously especially for the open-set and few-shot scenarios. Finally, the constructed CPPs are adopted as the input features of our proposed new TASE-Net to recognize the RF signals (the UAV models). Monte Carlo simulation results demonstrate that our proposed new UAV recognition approach using the TASE-Net can greatly outperform other existing deep-learning methods for open-set few-shot UAV identification.

Nowadays, wireless local-area networks (WLANs) are widely deployed in residential and commercial areas. The coverage quality is essential to users. The full coverage appears to be one of the most crucial problems to be considered during the network and access-point/router deployment (placement). We formulate the light-of-sight (LoS) coverage problem using the visibility-graph framework. In this work, for arbitrary multiply-connected or simply-connected polygonal/prismatic fields-of-interest subject to an arbitrary link-range restriction, we investigate how the full LoS coverage can be achieved by a minimum number of access-points/routers. Based on the new mathematical lemmas we derive, we design a novel graph-theoretical approach accordingly. Our proposed new scheme can be deemed the first-ever systematic approach to the best of our knowledge. Our proposed new approach is also evaluated in terms of the coverage efficiency, the number of access-points/routers, and the peak link-distance ratio for full LoS coverage in comparison with the existing solution to the art gallery problem.