通信时延与丢包下智能网联汽车控制性能分析

Analysis of Intelligent and Connected Vehicle Control Under Communication Delay and Packet Loss

网联协同控制是智能网联汽车的重要应用场景,而车联网的通信时延与丢包可能导致控制性能下降,甚至影响行车安全。为了分析时延与丢包对网联车辆控制的稳态与瞬态性能的影响,设计了网联控制器,并开展了仿真与实车试验。基于车辆动力学特性,将通信时延与丢包下的网联车辆控制分解为纵向控制与横向控制,进行了统一建模,并设计了控制器进行试验分析;搭建了网联自动驾驶的CarSim-Simulink联合仿真平台,及集成可模拟时延与丢包的LTE-V原理样机的智能网联汽车试验平台;开展了不同时延与丢包率下网联跟车控制与网联路径跟踪控制的仿真试验与实车试验。试验结果显示:时延与丢包对控制误差的影响形态有相似性;时延或丢包率取系统及工况参数有关的小值时,如试验中时延小于200ms或丢包率小于20%,工况随机因素对控制误差的影响可能超过时延与丢包的影响;在更大的时延或丢包率下,时延与丢包的出现方式(如出现时机等)对控制误差影响更大。研究结果表明:能实现针对网联车辆控制系统通信特性的控制器优化设计,使得当时延与丢包在工况相关阈值内时,系统控制误差有界。所揭露的规律一方面可用于对造成危险控制误差的时延与丢包工况进行预警,另一方面也可用于基于给定的稳态或瞬态控制误差边界,判定对应工况允许的时延与丢包率边界。

Connected cooperative control is an important application scenario of intelligent and connected vehicles (ICVs).However,communication delay and packet loss may impair control performance,and even influence driving safety.In this study,to analyze the influences of delay and packet loss on connected vehicle control performance in steady and transient states,connected controllers were designed,and simulation and field tests were conducted.The problem of connected vehicle control under delay and packet loss was divided into longitudinal and lateral controls based on vehicle dynamics,and these were modeled by a unified method.Controllers were then designed for empirical analysis.The CarSim-Simulink joint simulation platform for connected vehicle control and ICV testbeds using LTE-V prototypes that can simulate delay and packet loss were built,and joint simulation tests and field tests were conducted under various delays and packet loss rates.The test results indicate the following: The patterns of influences of delay and packet loss on control errors are similar.When the delay or packet loss rate is less than a certain value (e.g.,200 ms delay or 20% packet loss rate in the tests),the stochastic factors of the scenario might affect control errors more than delay or packet loss.Under a considerable delay or packet loss rate,the pattern of delay or packet loss (e.g.,time of occurrence),has a greater influence on control errors.This study also reveals that controllers can be optimized based on the communication characteristics of a connected vehicle control system to ensure bounded control errors when the delay or packet loss rate is within a scenario-related threshold.The revealed principles can be used for the warning of the delay and packet loss rate that can lead to dangerous control errors.They can also be used to determine scenario-related allowable bound of delay and packet loss rate,based on given bounds of steady or transient state control errors.