Sensing-Assisted Predictive Beamforming With Multipath Echo Signals

This paper investigates sensing-assisted predictive beamforming for vehicle-to-infrastructure (V2I) communication by exploiting integrated sensing and communication (ISAC) functionalities at the base station (BS). In the existing ISAC framework, the vehicle is usually treated as a point target and only the backscattered echo is used for sensing. However, in practice, the vehicle is an extended target composed of different reflecting/scattering surfaces, and hence the BS may receive multiple echoes. In this paper, by exploiting the directive scattering (DS) model for the incident signal at the vehicle surface, we demonstrate that the multipath echoes are substantial or even stronger, as compared with the backscattered echo. To capitalize on this observation, we develop a sensing-assisted predictive beamforming scheme with multipath echoes, where both the backscattered echo and multipath echoes are utilized for vehicle tracking. To accurately estimate the motion parameters of vehicles, we propose a particle filtering based tracking algorithm where the likelihood probabilities are designed based on the geometrical relationships of the multipath echoes. The beamforming vectors are then designed based on the predicted angles for establishing the communication links. Moreover, we present a particle swarm optimization (PSO) based algorithm to estimate the locations of the roadside scatterers when the prior information is unavailable. Extensive simulations are conducted to verify the superiorities of the proposed scheme.