基于免疫遗传模糊神经滑模的地面作战机器人轨迹跟踪控制  

A Trajectory Tracking Control Based on Immune Genetic Fuzzy Neural Sliding Mode for Ground Combat Robot

作　　者：李郁峰[1] 李魁武[2] 潘玉田[1] 郭保全[1] 

机构地区：[1]中北大学计算机与控制工程学院,山西太原030051 [2]西北机电工程研究所,陕西咸阳712099

出　　处：《火炮发射与控制学报》2016年第2期14-18,共5页Journal of Gun Launch & Control

基　　金：国家自然科学基金资助项目(51275489)

摘　　要：提出了一种基于运动学与动力学模型的免疫遗传模糊神经滑模混合控制器用于地面作战机器人的控制。算法中利用径向基神经网络逼近滑模控制的等效部分,并通过免疫遗传算法对径向基神经网络参数进行了优化,滑模控制的增益通过模糊控制策略进行了调节。利用该算法对圆形轨迹进行了跟踪控制仿真及试验分析,与传统的滑模控制相比,该算法能够有效克服系统的不确定性因素的影响,有效抑制了滑模控制中抖振现象,系统的动态轨迹跟踪性能得到了优化。Based on a kinematic controller and a sliding mode dynamics controller with immune gene-tic algorithm fuzzy neural network,a hybrid control algorithm is presented for ground combat robot. In this algorithm,the equivalent sliding mode control is approached by using radial basis function neural network (RBFNN),and the structure and parameters of RBFNN nonlinear hidden layers are tuned on-line according to immune genetic algorithm for the optimization of the parameters of networks with the gain of sliding mode control adjusted through the adaptive fuzzy systems. Simulation and experimental analysis of the circular trajectory tracking control are carried out by using this algorithm. Compared with the traditional sliding mode control,the proposed controller effectively overcomes the influences of uncertain factors,eliminates chattering of the sliding mode control,and optimizes the dynamic trajecto-ry tracking performance of the system.

关 键 词：滑模控制 神经网络 免疫遗传 模糊控制 轨迹跟踪控制 

分 类 号：TJ303.3[兵器科学与技术—火炮、自动武器与弹药工程] TP242[自动化与计算机技术—检测技术与自动化装置]