基于ANFIS的弹药自动装填机械手轨迹跟踪  

Trajectory Planning Simulation of Ammunition Auto-Loading Manipulator Based on ANFIS

作　　者：李郁峰[1] 潘玉田[1] 李魁武[2] 李科蕾[1] 

机构地区：[1]中北大学信息与通信工程学院,山西太原030051 [2]西北机电工程研究所,陕西咸阳712099

出　　处：《中北大学学报（自然科学版）》2012年第6期649-654,671,共7页Journal of North University of China(Natural Science Edition)

基　　金：高等学校博士学科点专项科研基金资助项目(20110142130010)

摘　　要：针对大口径自行火炮弹药自动装填机械手不确定性的复杂非线性系统,提出了一种基于自适应神经模糊推理(Adaptive Neuron-Fuzzy Inference System,ANFIS)的方法,对弹药自动装填机械手进行了运动学分析及轨迹跟踪仿真.该方法利用神经网络的学习机制补偿了模糊控制系统原有的缺点,前提参数采用误差反向传播法,结论参数采用最小二乘法.最后在Matlab中对弹药自动装填机械手进行了建模仿真.仿真与实验结果表明:该方法收敛速度快,泛化能力强,提高了弹药自动装填机械手的轨迹跟踪精度,能够满足实际要求,为进一步实现弹药自动装填机械手自适应控制打下基础.To the uncertain,complex and nonlinear ammunition auto-loading manipulator system for a large-caliber self-propelled gun,the modeling approach based on adaptive neuron-fuzzy inference system(ANFIS) for kinematics analysis and trajectory tracking simulation of ammunition auto-loading manipulator was presented.This method by using the neural networks′ learning mechanism compensated the original shortcomings of fuzzy control system.The premise parameters were adjusted by error back-propagation method and the conclusion parameters by least square procedure.The Matlab simulation and experimentation results of ammunition auto-loading manipulator indicate that the method has a strong generalization ability and fast convergence,and can improve the accuracy of trajectory tracking and meet the practical requirements.It lays the foundations for further realization of ammunition auto-loading manipulator adaptive control.

关 键 词：自动装填机械手 运动学分析 自适应神经模糊推理 MATLAB仿真 

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]