电磁直驱驾驶机器人模糊神经网络车速控制方法及试验验证

发布时间：2018-03-06 08:25

本文选题：汽车试验　切入点：电磁直驱　出处：《科学通报》2017年30期 　论文类型：期刊论文

【摘要】：为了实现不同驾驶循环工况的精确车速跟踪,提出了一种电磁直驱驾驶机器人模糊神经网络车速控制方法.通过电磁直线作动器直接驱动油门机械腿、制动机械腿、离合器机械腿和换挡机械手.给出了电磁直驱驾驶机器人控制系统结构和协调运动控制模型.在此基础上,设计了电磁直驱驾驶机器人模糊神经网络车速控制模型,电磁直驱驾驶机器人油门、制动、离合器机械腿和换挡机械手的位移作为网络模型的输入,试验车辆的车速作为网络模型的输出.输入变量的隶属度函数个数都为3,隶属度函数的类型都采用广义钟形函数,网络训练算法选用反向传播算法和最小二乘法相结合的混合学习算法.本文提出的方法与其他控制方法和人类驾驶员的驾驶性能进行了对比分析,试验验证了提出方法的有效性.
[Abstract]:In order to achieve accurate speed tracking under different driving cycle conditions, a fuzzy neural network speed control method for electromagnetic direct-drive robot is proposed. The accelerator is directly used to drive accelerator leg and brake mechanical leg. The structure of the control system and the coordinated motion control model of the electromagnetic direct drive robot are presented. On the basis of this, the fuzzy neural network speed control model of the electromagnetic direct drive robot is designed. The displacement of the electromagnetic direct drive robot throttle, brake, clutch mechanical leg and shift manipulator is used as the input of the network model. The speed of the vehicle is the output of the network model. The number of membership function of input variable is 3, and the type of membership function is generalized bell function. The network training algorithm is a hybrid learning algorithm which combines backpropagation algorithm and least square method. The proposed method is compared with other control methods and the driving performance of human drivers. The effectiveness of the proposed method is verified by experiments.
【作者单位】：南京理工大学机械工程学院;东南大学仪器科学与工程学院;
【基金】：国家自然科学基金(51675281,51205208) 江苏省六大人才高峰计划(2015-JXQC-003) 江苏省产学研前瞻性联合研究项目(BY2015004-02) 中央高校基本科研业务费专项(30916011302,NJ20160037)资助
【分类号】：TP242

【相似文献】