In order to realize the next generation of interconnected autonomous vehicles, integrated sensing and communication (ISAC) technology is applied to future Vehicle-to-Everything (V2X) networks. This paper proposes a beam tracking prediction scheme for radar-assisted communication based on neural network methods. Specifically, we consider an expanded vehicle model, focusing on the problem of communication beam alignment between Road Side Units (RSUs) and highly dynamic vehicles. In this regard, we propose a prediction beamforming method based on BFP-Net, which utilizes echo signals at the current moment to predict the optimal beam at the next moment. Simulation results show that the proposed method can effectively reduce system overhead and achieve higher beam gain.
BFP-Net:A Deep Learning Solution for Beamforming Prediction in Extended Vehicular Scenario based ISAC System
10.10.2023
1523618 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Vehicular Connectivity on Complex Trajectories: Roadway-Geometry Aware ISAC Beam-tracking
BASE | 2023
|