Inverse Kinematics of Compliant Manipulator Based on the Immune Genetic Algorithm

As the job of restaurant service robots calls for a smooth movement of the manipulator, controlling the posture of the manipulator is necessary in order to make a manipulator compliant status. So the problem of manipulator inverse kinematics has become particularly important. In order to avoid the traditional methods cumbersome formulization, and aiming at the deficiencies of BP algorithm in the training of neural networks, this paper presents an inverse kinematics solutions based on immune genetic algorithm, with inverse kinematics process being converted into the weight training problem of neural network. Experimental results show that, provided that the training samples are correctly chosen, the method used to solve manipulator inverse kinematics equation is practically feasible, for its high convergence speed and high accuracy. And it meets the real-time requirements.