基于GS-EKF滤波方法的车辆状态参数估计  

Vehicle State Parameter Estimation Based on Graded Series Extended Kalman Filter Method

作　　者：陈龙 刘巧斌[3] 陶磊 CHEN Long;LIU Qiao-bin;TAO Lei(College of Mechanical and Vehicle Engineering,Taiyuan University of Technology,Taiyuan 030024,China;State Key Laboratory of Automotive Simulation and Control,Jilin University,Changchun 130022,China;School of Vehicle and Mobility,Tsinghua University,Beijing 100084,China)

机构地区：[1]太原理工大学机械与运载工程学院,山西太原030024 [2]吉林大学汽车仿真与控制国家重点实验室,吉林长春130025 [3]清华大学车辆与运载学院,北京100084

出　　处：《东北大学学报（自然科学版）》2023年第8期1144-1151,共8页Journal of Northeastern University(Natural Science)

基　　金：国家重点研发计划项目(2020YFB1314001);汽车仿真与控制国家重点实验室开放基金资助项目(20210218);山西省基础研究计划(自由探索类)项目(20210302124119);山西省高等学校科技创新计划项目 (2021L085).

摘　　要：准确、高效的车辆状态估计是智能汽车实现精准控制的前提,因此迫切需要开展又快又准的状态估计算法研究.为此,提出一种分级串联型扩展卡尔曼滤波(graded series extended Kalman filter,GS-EKF)车辆状态参数估计方法,旨在保证估计精度的同时,提升算法的计算效率和鲁棒性.首先,基于分级串联思想,将初级扩展卡尔曼滤波估计的结果,作为次级状态估计的可量测控制输入信号,实现分级串联状态估计;然后,建立3自由度非线性动力学车辆状态参数估计模型,以方向盘转角及纵向、侧向加速度为输入变量和观测变量;最后,搭建联合仿真验证平台,对比分析4种不同算法的精度、鲁棒性以及效率.结果表明所提出的算法在精度和鲁棒性方面可达到粒子滤波的效果,而效率比粒子滤波提升了41.2%.Accurate and efficient vehicle state estimation is the prerequisite for intelligent vehicles to achieve precise control,thus there is an urgent need to carry out research on fast and accurate state estimation algorithms.To this end,a graded series extended Kalman filter(GS-EKF)method of vehicle state parameter estimation is proposed,which aims to improve the computational efficiency and robustness of the algorithm while ensuring the estimation accuracy.Firstly,based on the graded series idea,the result of the primary extended Kalman filter estimation is used as the measurable control input signal of the secondary state estimation to realize the hierarchical cascading state estimation.Then,a 3-degree-of-freedom nonlinear dynamic vehicle state parameter estimation is established which takes the steering wheel angle and longitudinal/lateral acceleration as input variables and observation variables.Finally,a joint simulation verification platform is built to compare and analyze the accuracy,robustness and efficiency of four different algorithms.The results show that the proposed algorithm can achieve the effect of particle filter in terms of accuracy and robustness,and the efficiency is 41.2%higher than that of particle filter.

关 键 词：智能汽车 状态估计 分级串联 扩展卡尔曼 鲁棒性 

分 类 号：U461.1[机械工程—车辆工程]