Self-learning model predictive control method for assembly of robot electronic components

Abstract

The invention discloses a self-learning model predictive control method for assembly of robot electronic components. The self-learning model predictive control method includes the steps that the difference between the position, speed and acceleration output of an end effector and a joint and an expected value is input in real time; a model predictive controller is obtained through self-learning training, and the controlling torque delay difference is output; delta[tau] is further used as feedback and as the input of the assembly robot of the electronic components so as to output the actualtarget position, speed and acceleration of the robot joint and the robot end effector in the assembly process of the electronic components; the difference between the above and the expected value is fed back to a model predictive controller to realize the self-learning of the predictive controller; compared with an existing robot assembly control method, according to the self-learning model predictive control method, the self-learning model predictive controller with the difference as input and output has the characteristics of high speed and high precision, and the needs of high precision on-line assembly of the electronic components under various interference factors can be met.

Description

technical field [0001] The invention is applicable to the field of robot intelligent assembly, and in particular relates to a self-learning model predictive control method for robot electronic component assembly. Background technique [0002] Assembly robot is a multi-input, multi-output nonlinear, strongly coupled dynamical system with time-varying position. The movement of each joint is affected by other joints, and the inertial load acting on each joint varies in a wide range with the shape of the manipulator arm. In order to solve the uncertainty of the environment and the uncertainty of the dynamic characteristics of the robot In order to achieve precise assembly control, it is necessary to design a corresponding efficient control algorithm for the dynamics of the robot and its various uncertainties, and apply it to the actual control process. [0003] The robot has skillful assembly actions such as clamping, pressing, inserting, rotating, and pushing. The completion o...

AI Technical Summary

Problems solved by technology

[0004] Therefore, due to the small size and strong fragility of electronic components, they are easily damaged during the assembly process. In order to ensure safer and more efficient assembly of electronic components, higher precision robot electronic component assembly is required.

Images