基于感通算融合和信息年龄优化的车联网多节点协同感知

Abstract

<p>面向未来自动驾驶系统中的实时性业务需求（如高清地图更新），基于感知-通信-计算融合，引入信息年龄作为实时性度量，设计感通算融合的车联网多节点协同感知机制。在通信-计算资源和车辆能耗约束下，优化调度感知节点信息采集和传输处理，最小化感知信息的平均信息年龄；提出基于李雅普诺夫的在线调度算法，将复杂的长期随机优化问题转化为单时隙在线优化问题，并设计低复杂度算法求解。仿真表明，与现有仅考虑通信与计算融合的机制相比，所提机制信息实时性可提高9%～50%。<br></p>

For the requirements of real-time services in autonomous driving systems, such as high-definition (HD) maps, based on the integration of sensing, communication, and computing, a multi-node cooperative sensing mechanism was proposed with the age of information (AoI) as the real-time indicator. Considering the constraints on communication and computing resources and vehicle energy consumption, the information collection, transmission and processing of sensing nodes were optimally scheduled to minimize the AoI averaged over time. A Lyapunov-based online scheduling algorithm was proposed to transform the long-term stochastic optimization problem into an online optimization problem, which could be solved with low complexity. Compared with the existing mechanism considering integrated communication and computing, the proposed mechanism can improve real-time performance by 9%~50%.