The invention discloses a
finite time neural network optimization method for solving the
inverse kinematics of a redundant
manipulator. The
finite time neural network optimization method comprises thefollowing steps that 1), an expected target track r*(t) and a joint angel theta*(0) expected to be returned of an end
effector of the redundant
manipulator are determined, and the end
effector of the
manipulator is deviated from the position of the expected track; 2), final state attraction optimization indexed are designed, a quadratic
programming scheme based on the final state attraction is constructed, wherein an initial
joint angle of the redundant manipulator during actual movement can be arbitrarily designated, the initial
joint angle theta(0) of the redundant manipulator during actual movement is given, theta(0) is taken as a motion starting point, and the formed repeated motion
programming scheme is described as the quadratic
programming with the final state
attractor optimization indexes; 3) a final state neural
network model of a finite value
activation function is constructed, and a finite value final state neural network is used for solving a time-varying matrix equation; and 4), the result which is obtained by solving the equation is used for controlling each joint motor to drive the manipulator to execute tasks. The
finite time neural network optimization method has the advantages of being high in precision and capable of converging in finite time.