Distributed Multi-View Environment Sensing in Wireless Communication Networks

Artikel i vetenskaplig tidskrift, 2026

In this paper, we propose a generic distributed multi-view sensing framework to deal with the three crucial challenges, i.e., physics-inspired signal propagation modeling, large-scale region of interest (ROI) pixelation, and distributed sensing architecture and algorithm design, in typical ISAC (integrated sensing and communication)-enabled wireless networks. First, we propose a unique pixel coordinate system with non-uniform and affine pixel division and mapping, which helps to resolve the occlusion effects between pixels and can be well tailored for common computational imaging technology. Second, inspired by Huygens’ principle, we propose a novel integral-form signal propagation model that well captures the overall amplitude and phase effects of the pixels with arbitrary shapes and sizes, thus leading to a greatly reduced quantization error and making it suitable for large-scale scenario sensing. The proposed system model allows free pairing of transmitters and receivers in the network and makes full use of the multi-view observations, further resolving the occlusion problem any individual receiver may encounter in large-scale environment sensing. We then propose a low computational complexity alternating optimization (AO) algorithm, which iteratively performs sparse reconstruction and occlusion detection to obtain accurate sensing results. Finally, the performance of the proposed method is analyzed. Extensive numerical results verify the effectiveness of the proposed method.