基于神经网络的冗余度TT-VGT机器人的运动学求解

THE KINEMATICS OF REDUNDANT TT VGT MANIPULATORS BASED ON NEURAL NETWORK

应用ＢＰ神经网络对冗余度ＴＴ－ＶＧＴ机器人的位姿正解进行训练学习，进而求解机器人的位姿反解问题．根据网络模型求得机器人的一、二阶影响系数，应用神经网络求解雅可比矩阵的伪逆．

The forward displacement analysis problem of TT VGT manipulators with redundant degree of freedom is trained based on BP neural network, and then a solution to inverse displacement analysis problem is obtained. According to the above the network model, the first and the second order influence coefficients are derived, and the pseudo inverse of Jacobian matrix is obtained by using neural network. Finally the simulation calculation of kinematics for a seven celled tetrahedron tetrahedron variable geometry truss manipulator is given for illustration.