Predictive Beam Tracking with Cooperative Sensing for Vehicle-to-Infrastructure Communications

Abstract

Beam alignment is a critical issue for millimeter wave (mmWave) communication in high-mobility vehicular scenarios. In order to enhance the beam alignment performance, in this article, we investigate a dual-functional radar-communication system where the intelligent vehicle can actively cooperate with the roadside stations by sharing its sensing results. Based on the state evolution model of the vehicle, an Extended Kalman filter for beam tracking is employed. It is shown that, with the radar reflections as well as the sensing results from the vehicle, the proposed scheme can track the mobility of the vehicle and predict the beam directions more accurately, which can benefit the communication. The cooperative sensing between the vehicle and the roadside stations can be used to predict beam directions with low overhead for vehicles in complex scenarios, such as curves or crossovers. Simulations demonstrate that the proposed scheme achieves better beam tracking performance compared to the conventional pilot-based ones.