Mechanical arm motion planning method based on fixed parameter neural network

Abstract

The invention discloses a mechanical arm motion planning method based on a fixed parameter neural network. The method comprises the following steps of S1, establishing an inverse kinematics equation of a mechanical arm according to the Jacobian matrix and a preset target track at the tail end of the mechanical arm; S2, establishing physical limit double-end inequality constraint of the mechanicalarm according to actual joint physical limit constraint parameters of the mechanical arm; S3, formulating the inverse kinematics equation and the physical limit double-end inequality constraint into atime-varying quadratic programming problem; S4, designing the fixed parameter neural network with a penalty function to solve a time-varying quadratic programming problem; and S5, transmitting the obtained angle information of the mechanical arm to a lower computer controller of the mechanical arm, and driving the mechanical arm by the lower computer controller to move to complete the target track tracking task. According to the method, the time-varying quadratic programming problem is solved by adopting the novel fixed parameter neural network with the penalty function, the convergence speedis faster, the calculation precision is higher, and the inequality constraint condition in a quadratic programming scheme can be effectively solved.

Description

technical field [0001] The present invention relates to the technical field of mechanical arm movement subject to various constraints, in particular to a mechanical arm motion planning method based on a solid parameter neural network. Background technique [0002] The repetitive movement of the robotic arm is an extremely common movement in industrial production. Repeated movement requires that each joint of the manipulator can return to its initial state after completing a period of closed end trajectory movement. In this case, it can ensure that the initial state of each period of the manipulator is consistent. If the joints of the robotic arm do not all return to their initial state after completing a cycle of motion, it is considered that there is joint deviation. If the joint deviation occurs during the periodic motion, the accuracy of the motion control of the manipulator will decrease, or an additional reset adjustment process will be required, resulting in a serious...

AI Technical Summary

Problems solved by technology

However, LVI-PDNN cannot achieve exponential convergence speed. In order to make full use of the characteristics of the time-varying system of the robot's motion planning problem, Zhang et al. designed a variable parameter convergence differential neural network with super-exponential convergence effect. [2] , and apply the network to a repetitive motion planning scheme for redundant robots

However, the variable parameter convergence differential neural network cannot effectively solve the inequality constraint problems in the motion planning scheme of the manipulator, such as: joint limit physical constraints

Images