Integrated Sensing, Communication, and Computation Over the Air: Beampattern Design for Wireless Sensor Networks

Abstract

In the future sixth-generation wireless communications, wireless sensor networks (WSNs) are expected to support target sensing, information communication, and computational tasks concurrently over the same spectrum. Integrated sensing and communication (ISAC) and over-the-air computation (AirComp) arise as two promising techniques. In this article, we design an integrated sensing, communication, and computing framework to improve spectrum efficiency and quality of service in WSNs. We investigate omnidirectional and directional beampattern designs to minimize AirComp error. Leveraging the derived directional patterns, we further consider tradeoff beampatterns to balance sensing and AirComp performance under power constraints. A mismatch-based design is formulated to minimize AirComp errors. To solve these nonconvex problems, we propose an alternating optimization approach using singular value decomposition and projection techniques to jointly optimize sensor and access point beampatterns. In particular, we propose a low-complexity two-step projection-based gradient descent method to solve the convex problems with two norm-ball constraints. In addition, convergence and complexity analysis of the proposed methods is provided. Simulations demonstrate the performance of the proposed beampattern designs.