基于蚁群神经网络的机械手自组织逆运动规划

Self-organizing Inverse Kinematics Planning for Manipulators Based on Ant Colony Neural Network

将蚁群神经网络应用于机械手的自组织逆运动规划,通过设置目标函数将神经网络的训练问题转化为在连续域内求极小值问题。在算法前期,仅需要取得运动路径的采样点集,不需要先期求解采样点的运动学逆解,算法实现了机械手的自组织逆运动规划,采用连续蚁群算法求解该连续域函数极小值问题。实验表明算法有较高的控制精度与收敛速度。

Ant colony neural network is used to solve self-organizing inverse kinematics planning for manipulators. The neural network training problem is translated into the problem of searching the minimum value of the function in a continuous space by setting the objective function. In the early stage of the algorithm ,the sample point set of the motion route is only needed to be obtained for training the neural network,and the inverse kinematics solution is not needed. The algorithm achieves self-organizing manipulator inverse kinematics planning for the neural network. The continuous ant colony algorithm is used to solve the problem of searching the minimum value for the continuous function. Experimental resuits show that the control algorithm has higher accuracy and convergence speed.