Biped robot gravity center control method and simulation system based on reinforcement learning

Abstract

The invention discloses a biped robot gravity center control method and system based on reinforcement learning, equipment and a storage medium. The method comprises the steps that a robot three-dimensional assembly construction step is conducted, specifically, a robot three-dimensional model is constructed on the basis of basic cube assemblies, and the freedom degree of each part is set; a robot gravity system construction step is conducted, specifically, parameters of the basic cube assemblies and the gravity center of a robot in an initial state are calibrated, and a robot gravity system is constructed; a robot inertia system construction step is conducted, specifically, a rotational inertia model based on joints of a leg is constructed; a robot gravity center control step is conducted, specifically, a gravity center control strategy of the gravity center and steady-state walking of the biped robot is constructed, and calibration and setting of gravity center control related parameters are conducted; and a robot analog simulation step is conducted, specifically, control simulation of the gravity center and steady-state walking of the biped robot is achieved on the basis of a reinforcement learning framework. According to the method, the efficient and instant gravity center control of the robot can be realized, and the robot adapts to changeable road environments.

Description

technical field [0001] The invention belongs to the fields of robots and artificial intelligence, and in particular relates to a method for controlling the center of gravity of a biped robot based on reinforcement learning and a simulation system. Background technique [0002] In recent years, with the improvement of machinery manufacturing technology and artificial intelligence technology, the demand and application fields of robots have been greatly expanded. Various industrial robots, sweeping robots, service robots, etc. have been successfully applied in various industries. In the field of service robots, bipedal high-simulation robots have always been the most difficult and most important research direction. One very important reason is that biped robots need efficient and real-time center of gravity control during dynamic operation to ensure rapid and stable of the mobile. [0003] However, due to the comprehensive influence of the current mechanical manufacturing lev...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the present invention proposes a method and system for controlling the center of gravity of a biped robot based on reinforcement learning, which is used to solve the problem that the center of gravity of a biped robot is not effective and cannot adapt to the changing road environment in real time

Images