基于改进的粒子群优化算法的轮胎参数辨识  

Parameter Identification of the Tire Model Based on an Improved Partical Swarm Optimization Algorithm

作　　者：宋晓琳[1,2] 李红[1,2] 郭孔辉[1,2] 

机构地区：[1]湖南大学 [2]汽车车身先进设计制造国家重点实验室,长沙410082

出　　处：《科技导报》2011年第9期53-56,共4页Science & Technology Review

基　　金：湖南大学汽车车身先进设计制造国家重点实验室自主研究课题(61075004);中国高水平汽车自主创新能力建设项目

摘　　要：轮胎是汽车的重要组成部分,其特性分析是研究汽车动力学的基础,其模型的精度直接影响整车模型仿真的精度,多采用粒子群优化算法对轮胎参数进行辨识。参考自然界生物进化现象,在基本粒子群算法的基础上提出带变异阀值的多种群粒子群算法。该算法采用多个种群同时进化以保证粒子群的多样性,同时可改善全局收敛的可靠性,采用变异阀值可避免优化算法陷于局部收敛现象的发生。将该方法应用于轮胎参数辨识,并与其他优化算法辨识结果进行比较,该方法结果能够更好地与实验数据吻合,证明该方法辨识精度高,在轮胎参数辨识中有较好的应用性。Tire is an important part of the vehicle.The behaviour of the tire is of basic importance to the vehicle dynamics,and plays a vital role in vehicle＇s performance,so the precision of a tire model affects the simulation reliability of the whole vehicle model.Partical swarm optimization algorithm is used to identify the tire model parameters in this paper.According to the organic evolution in nature,the multi-population with a variation threshold partical swarm optimization algorithm is proposed to keep the population diversity and improve the reliability of holistic convergence.The variation threshold is to avoid the problem of converging to a part-optimum.Comparing the simulation results with the test results,it is shown that the simulation data from the multi-population with a variation threshold partical swarm optimization algorithm would fit the test data better than those from other optimization algorithms,the identification accuracy is slightly higher.The multi-population with variation threshold optimization algorithm has a good application prospect in tire parameter identification.

关 键 词：轮胎 参数辨识 多种群 变异 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构] U463.341[自动化与计算机技术—计算机科学与技术]