Hexapod robot impedance control method based on reinforcement learning

Abstract

The invention discloses a hexapod robot impedance control method based on reinforcement learning. The method comprises the following steps of establishing a hexapod robot dynamic system which has a noise parameter and is based on dynamic motion primitives; determining a torque control expression based on impedance control; determining a table form of a variable gain table; determining a cost function of a control system; and determining a parameter updating rule based on a path integral learning algorithm. According to the control method, system parameters are learned and updated by means of the path integral learning algorithm, so that a value of the cost function is as small as possible, and a robot can continuously adjust a reference track of foot end motion and a gain of a controller under the interference of an uncertain force field to obtain a good variable impedance control effect, and can move to an ideal target point in an expected form.

Description

technical field [0001] The invention relates to the field of robot control and reinforcement learning, in particular to a method for controlling the impedance of a hexapod robot based on reinforcement learning. Background technique [0002] In the field of hexapod robot control, the control goal is usually the stable movement of the robot foot according to a given desired trajectory, and the controller can reduce the error between the expected rotation angle and the actual rotation angle of the joint through position control. However, in the uneven and complex ground environment, the foot end of the hexapod robot may be unstable due to uneven force, so it is difficult to achieve the purpose of compliant control only by using position control. [0003] Impedance control is one of the most widely used methods in the compliance control of hexapod robots. It makes the position and force satisfy the desired dynamic equation by changing the damping and stiffness of the end effecto...

AI Technical Summary

Problems solved by technology

However, the traditional impedance control has the following disadvantages: the control parameters are fixed, and it is difficult to deal with the nonlinear time-varying interference in the unstructured environment

For example, Li Zhengyi et al. proposed a neural network-based impedance control algorithm in the paper "Robot Impedance Control Method Adapting to Unknown or Changing Environmental Stiffness and Damping Parameters", which enables the robot to have the ability to change impedance, but the neural network method has the following Disadvantages: First, it is necessary to establish a relatively complex network model; second, it is necessary to calculate the gradient and complete the backward propagation, which requires a large amount of calculation

Images