E-resources
Peer reviewed
Open access
-
Mao, Liangan; Li, Li
Journal of physics. Conference series, 12/2022, Volume: 2400, Issue: 1Journal Article
In view of the complexity of the traditional inverse kinematics solution, it is solved by the mapping of the BP neural network, but the precision is low and the network performance is not good. By using FGA to optimize BP neural network, the shortcomings of the neural network itself can be overcome. Taking the pose matrix and joint angle of the manipulator as the input and output variables of the prediction network, the inverse solution prediction BP network of the manipulator is constructed, and then samples are selected for training. Finally, the network test and simulation results show that the optimized BP neural network has higher prediction precision and improved network performance.
