Physics

Computer network

Beamforming

Wireless

Communications system

Engineering

Telecommunications link

Artificial intelligence

Kalman filter

Throughput

Extended Kalman filter

Electronic engineering

Computer science

Real-time computing

Kinematics

Radar

Classical mechanics

Telecommunications

IEEE Transactions on Wireless Communications

In vehicular networks of the future, sensing and communication functionalities will be intertwined. In this article, we investigate a radar-assisted predictive beamforming design for vehicle-to-infrastructure (V2I) communication by exploiting the dual-functional radar-communication (DFRC) technique. Aiming for realizing joint sensing and communication functionalities at road side units (RSUs), we present a novel extended Kalman filtering (EKF) framework to track and predict kinematic parameters of each vehicle. By exploiting the radar functionality of the RSU we show that the communication beam tracking overheads can be drastically reduced. To improve the sensing accuracy while guaranteeing the downlink communication sum-rate, we further propose a power allocation scheme for multiple vehicles. Numerical results have shown that the proposed DFRC based beam tracking approach significantly outperforms the communication-only feedback based technique in the tracking performance. Furthermore, the designed power allocation method is able to achieve a favorable performance trade-off between sensing and communication.

Radar-Assisted Predictive Beamforming for Vehicular Links: Communication Served by Sensing