Predictive Beam Tracking and Power Allocation With Cooperative Sensing for V2I Communication

Integrated sensing and communication (ISAC) technology enables sensing-assisted communication in high-mobility vehicle-to-infrastructure (V2I) networks via beam tracking and vehicle state prediction. In this paper, from the practical perspective, we propose a cooperative sensing-aided V2I communication framework for more realistic scenarios that involve non-linear trajectories and multi-user interference (MUI). To accurately track the vehicle, a beam tracking algorithm is designed that jointly processes vehicular measurement and infrastructure sensing information on azimuth angle, distance, and velocity to predict the vehicle state. Based on this framework, the closed-form communication throughput and sensing posterior Cramér-Rao bound (PCRB) are analyzed to provide explicit performance evaluation. In addition, a power allocation scheme among users is designed to maximize the sum communication rate while guaranteeing sensing performance in terms of PCRB, considering the performance degradation and coupling effect caused by MUI. The complementary geometric programming problem is solved by successively refined geometric programming problems with convex approximation. Numerical results demonstrate the superiority of the proposed V2I communication framework compared with existing schemes, which improves communication performance by optimizing power allocation and mitigating MUI with more accurate vehicle state prediction.