Industrial mechanical arm motion planning method based on reinforcement learning algorithm

Abstract

The invention discloses an industrial mechanical arm motion planning method based on a reinforcement learning algorithm, and belongs to the field of application of reinforcement learning to mechanical arm motion planning. The reinforcement learning Actor-Critic algorithm is applied to the motion planning of a mechanical arm, so that an interactive relationship is established between the mechanical arm and the environment, and the adaptive capacity of the mechanical arm to the environment is improved through real-time interaction with the environment for training, so that autonomous learning control is realized; and firstly, a simulation environment of a mechanical arm hand-eye system is built, then a reinforcement learning algorithm model is built according to the simulation environment, finally, motion planning training of the mechanical arm is completed, and intelligent control over the mechanical arm is achieved. The mechanical arm motion planning algorithm based on reinforcement learning has good environmental adaptability and stability.

Description

technical field [0001] The invention belongs to the field of reinforcement learning applied to motion planning of a mechanical arm, and in particular relates to a motion planning method for an industrial mechanical arm based on a reinforcement learning algorithm. Background technique [0002] Motion planning of manipulators to accomplish specific complex tasks in an uncertain environment has always been a very challenging problem. Traditional control methods often rely heavily on the system model. However, the model often has the characteristics of high-order, nonlinear, multi-variable, and strong coupling. It is difficult to make the manipulator system have good adaptability and certain autonomy. In recent years, artificial intelligence technology has developed vigorously, which provides new ideas for autonomous learning and control of robotic arms. The core idea of ​​artificial intelligence is to introduce an online learning mechanism in the planning and control of the ma...

AI Technical Summary

Problems solved by technology

The value evaluation algorithm of reinforcement learning is based on the value iteration method, which is conducive to the convergence of the value function to the optimum but is not suitable for continuous motion processes. The policy evaluation algorithm based on parameter optimization is more suitable for high-dimensional and continuous action control, and has better convergence. attributes, but single-use policy evaluation is prone to fall into local optimum

Images