改进粒子滤波算法的车辆状态估计研究

Research on Vehicle State Estimation Based on Improved Particle Filter Algorithm

针对传统粒子滤波存在的粒子退化和样本枯竭问题,采用了一种基于模拟退火粒子群优化粒子滤波(SAPSO-PF)车辆状态估计算法。首先基于Dugoff轮胎模型建立了汽车7自由度车辆模型。然后利用SAPSO-PF算法通过对低成本传感器测量到的纵向加速度、侧向加速度、方向盘转角和各车轮轮速信号,准确估计车辆的纵向速度、侧向速度以及横摆角速度。并在Carsim和Matlab/simulink环境下以实车场地实验数据进行仿真验证。多工况下的仿真试验结果表明,运用SAPSO-PF算法对车辆纵向速度、侧向速度横摆角速度的估计有良好的准确性。

In view of the problem that in the particle degradation and sample depletion in conventional particle filters,the algorithm is propoesd to estimate the vehicle state based on Simulated Annealing Particle Swarm Optimization Particle Filter(SAPSO-PF).Firstly,the 7-DOF vehicle model was established based on the Dugoff tire model.Then use SAPSO-PF algorithm to accurately estimate the vehicle’s longitudinal velocity,lateral velocity and yaw rate by measuring the longitudinal acceleration,lateral acceleration,steering wheel angle and wheel speed signals measured by low-cost sensors.The algorithm is verified by simulation of actual vehicle field experimental data by Carsim and Matlab/simulink.The results of simulation under multiple operating conditions show that the SAPSO-PF algorithm can accurately estimate the vehicle longitudinal velocity and lateral velocity yaw rate.