Value Matters: A Novel Value of Information-Based Resource Scheduling Method for CAVs

The Internet of Vehicles (IoV) can support applications in connected autonomous vehicles (CAVs), the implementation of which can effectively improve traffic efficiency. However, safety-related CAV applications have very strict requirements on the reliability and latency of each packet, which is difficult to achieve due to limited resources and the high dynamics of CAVs. In this article, we investigate communication resource scheduling for remote autonomous driving (AD) to improve the performance of the remote control system when network resources are constrained. Specifically, we introduce a novel performance metric, i.e.,
value of information
(VoI), to capture how sending a packet will affect the performance of CAV driving safety and efficiency, i.e., the value of the packet on the considered CAV system. The formulation of VoI is derived using the Lyapunov optimization method, and the lower-bound for the performance of the AD system with a VoI-based scheduling strategy is analyzed. Then, a communication resource scheduling approach is proposed based on the VoI of each packet. Simulation results demonstrate that the proposed VoI-based resource scheduling approach is capable of accurately assessing the impact of information transfer on system performance, while ensuring the CAV's safety and enhancing traffic efficiency.