基于RBF神经网络车速控制的再生制动系统测试  

Test of regenerative braking system with speed tracking control based on RBF neural network

作　　者：陈燎[1] 谢明维 盘朝奉[1,2] 周孔亢[1] 

机构地区：[1]江苏大学汽车与交通工程学院,江苏镇江212013 [2]江苏大学汽车工程研究院,江苏镇江212013

出　　处：《江苏大学学报（自然科学版）》2016年第3期256-263,共8页Journal of Jiangsu University：Natural Science Edition

基　　金：国家自然科学基金资助项目(51105178;51475213);江苏省"六大人才高峰"项目(2013-XNY-002);山东省高等学校科技计划项目(J13LN38);江苏省自然科学基金资助项目(BK2011489)

摘　　要：针对纯电动汽车再生制动系统优劣评估问题,提出了一种基于车速自动跟踪的再生制动系统测试方法.首先根据离线测试数据推导车速跟踪开环控制动态数学模型;然后用RBF神经网络搭建车速跟踪闭环控制驾驶员模型;最后利用PSO算法对RBF神经网络参数进行优化.在试验室自主研发的整车惯性模拟台架上进行试验,试验结果表明:用RBF神经网络算法控制车速跟踪相比传统模糊PID控制减小了车速跟踪误差,提高了再生制动系统测试的准确性,该方法在实际再生制动测试中应用是可行的.To evaluate the regenerative braking system of electric vehicle, a testing method for regenerative braking system was proposed based on automatic speed tracking. According to the off-line data from vehicle, the dynamic mathematical model for open-loop control of speed tracking was established. The driver model of speed tracking closed-loop control was built by RBF neural network,and the parameters in RBF neutral network were trained by PSO algorithm. The test was conducted on vehicle inertia simulation bench. The experimental results show that compared with traditional fuzzy PID control method,the speed tracking error is reduced by the RBF neural network algorithm,and the accuracy of regenerative braking system test is also improved. The feasibility of the proposed method is verified in actual regenerative braking test.

关 键 词：再生制动测试 RBF神经网络 车速跟踪 循环工况测试 PSO算法 

分 类 号：U462.3[机械工程—车辆工程]