Task Offloading Delay Minimization with Sensing Quality Assurance in VEC-ISAC Network

To address the dual requirements of multi-target sensing and low-latency computation offloading in traffic environments, this paper proposes a vehicular edge computingintegrated sensing and communication (VEC-ISAC) network. In this network, vehicles offload tasks while simultaneously sensing other vehicles to ensure driving safety, and the reconfigurable intelligent surface (RIS) is deployed to enhance the communication link between vehicles and the base station (BS). In this paper, we formulate a joint vehicle beamforming design, RIS phase shift, and computation resource allocation problem to minimize task offloading delay while ensuring the target sensing quality. However, the optimization problem is non-convex, making it difficult for traditional methods to solve. To tackle this difficulty, we reformulate it into a Markov decision process (MDP) and solve it using the multi-agent twin delayed deep deterministic policy gradient (MATD3) algorithm. Simulation results demonstrate that the proposed approach significantly reduces offloading delay while ensuring the required sensing performance, outperforming other benchmark algorithms.