通感一体化辅助的车联网协同感知资源分配方案研究OA

To address the issues in existing cooperative perception schemes for internet of vehicles—namely,saturated or underutilized bandwidth and computing resources,large volumes of perception data transmission,and high latency caused by integrated sensing and communications(ISAC)designs that fail to balance sensing and communication performance—this paper proposes an ISAC-assisted resource allocation scheme for cooperative perception in internet of vehicles.To reduce perception data transmission and enable efficient utilization of bandwidth and computing resources,a relative-posi-tion-based multi-layer region-of-interest scheme for cooperative perception is proposed,in which connected autonomous vehicles(CAVs)perceive only regions of interest and select early or intermediate fusion based on task processing time.Furthermore,a time allocation scheme for sensing,communication,and computation is designed based on the time-division dynamic frame structure of ISAC-enabled devices.The impacts of different time allocation ratios on the mutual information(MI)of sensing and communication are analyzed,and time allocation is determined accordingly.An optimization problem is formulated with the objective of minimizing the completion latency of cooperative perception tasks and is solved using a hybrid action space multi-agent deep deterministic policy gradient(HAS-MADDPG)algorithm.Simulation results demon-strate that,compared with baseline schemes,the HAS-MADDPG algorithm can effectively reduce the execution latency of cooperative perception.