An online machine learning control method for planetary rover wheels

Abstract

The invention provides an online machine learning control method for planetary exploration vehicle wheels, which relates to the field of robot control technology, including: determining the pseudo-driven wheels of the planetary exploration vehicle; obtaining the expected moving speed of the planetary vehicle, the hook traction force and support force of the pseudo-driven wheels , torque, and the speed control value output by the PID controller; determine the driving speed of the pseudo-driven wheel according to the expected moving speed; input the speed control value, hook traction force, support force and torque into the online learning neural network model for online learning, and the online learning neural network The output of the network model is determined as the approximate value of the hook traction force; the PID control gain adjustment of the pseudo-driven wheel is determined according to the partial derivative relationship between the approximation value and the control coefficient of the PID controller; the update speed control is determined according to the PID control gain adjustment. The invention combines PID control with online learning neural network model, so that the driving control of the pseudo-driven wheel has stronger adaptability and stability.

Description

technical field [0001] The invention relates to the technical field of robot control, in particular to an online machine learning control method for planetary exploration vehicle wheels. Background technique [0002] In recent years, countries around the world have successively launched a series of planetary exploration missions. As an effective carrier platform for scientific detection instruments, planetary vehicles play a pivotal role in planetary exploration projects. In the process of carrying out the exploration mission of the planetary vehicle, not only the planetary vehicle is required to have basic functions such as forward, backward, turning, climbing, etc., but also has the ability to adapt to the terrain and certain obstacle avoidance capabilities. [0003] At present, most of the planetary vehicles that land on the surface of the moon, Mars and other celestial bodies to perform survey tasks are multi-wheeled mobile robots. In the face of complex and diverse ter...

AI Technical Summary

Problems solved by technology

However, for different terrain environments, PID control also exposes shortcomings such as poor adaptability, which leads to unsatisfactory control effects on planetary vehicles.

Images