运动学与动力学融合的车辆状态估计方法

李克强++罗禹贡++边明远++戴一凡++陈龙

摘 要：汽車状态参数的准确获得是保证汽车主动安全系统有效性的重要要求。分布式电驱动车辆的新型结构为传统基于动力学的状态参数估计方法的突破提供了可能。通过分析基于运动学与动力学方法各自不同的误差特性，该文提出了对两种估计方法的估计结果进行融合处理的分布式电驱动车辆状态参数估计方法。利用理论推导，证明了该方法将能够有效的提高不同工况下的估计精度，提高估计方法的工况适应性。为验证该方法的有效性，开发了CarSim与Simulink联合仿真试验平台。仿真结果表明，所提出的误差加权的融合状态观测方法提高了分布式电驱动车辆状态参数观测精度和鲁棒性。

关键词：分布式电驱动车辆 车辆状态估计 多方法融合

Vehicle State Estimation Based on Kinematic Model and Dynamic Model Merging

Li Keqiang Luo Yugong Bian Mingyuan Dai Yifan Chen Long

（Tsinghua University）

Abstract：Vehicle state parameters are essential for active safety control. Distributed electric vehicle with a new structure brings a breakthrough for the traditional dynamics state parameter estimation method. This paper presents a novel estimation method for distributed electric drive vehicle by analyzing different characteristics of the estimation errors of kinematics and dynamics estimation methods. This method merges the results of the two estimation methods with weighting coefficients. With a mathematical deduction， it shows that this method can effectively improve the estimation accuracy and applicability under different conditions. A CarSim and Simulink co-simulation test platform is developed to verify the effectiveness of the method. Simulation results show that the proposed method improves the state estimation accuracy and robustness of distributed electric drive vehicle state parameters.

Key Words：Distributed electric vehicle； Vehicle state estimation； Multi-method merging