A robot contact control method based on convex optimization method

Through the robot contact control method based on convex optimization method, the problem of insufficient stability and robustness in complex tasks and uncertain environments is solved, and the robot can work stably and achieve task requirements in uncertain environments.

CN115185180BActive Publication Date: 2025-03-18ROKAE (BEIJING) TECH CO LTD
2 Cites 0 Cited by

Patent Information

Application Number
CN202210738771.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-06-28
Publication Date
2025-03-18
Estimated Expiration
2042-06-28

AI Technical Summary

Technical Problem

Existing robot contact control technology is difficult to maintain stability and robustness when facing complex tasks and uncertain environments, resulting in the controller being unable to provide appropriate control gain, the system robustness is reduced, and the task requirements are not met.

Method used

The robot contact control method based on convex optimization method is adopted. By constructing a discrete time system dynamic model, performing Z transformation, obtaining the transfer function matrix expression, and performing parameterization of the controller transfer function, converting the controller solution problem into a standard convex optimization problem, and introducing an environmental model, numerical optimization method is used to solve the optimization problem, and obtaining the optimal controller under a closed-loop system.

Benefits of technology

When facing uncertain and unknown environments, the designed controller shows stronger stability and robustness. The robot can work steadily in contact working scenarios that are several times the hardness of the expected working conditions, and meet the task requirements. It is easy to deploy, simple to solve, does not require complex online calculations, and avoids cumbersome manual parameter adjustment.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The present invention proposes a robot contact control method based on a convex optimization method, comprising: step S1, setting the parameters of a discrete time invariant system dynamics model; step S2, performing a Z transform on the dynamics model to obtain a transfer function matrix expression of the system dynamics; step S3, obtaining a controller transfer function K(z) of an open-loop system and a controller transfer function T(z) of a closed-loop system under discrete time; step S4, converting T(z) into a linear combination form of multiple matrices by performing an affine parameterization on T(z); step S5, converting the controller solution problem into a standard convex optimization problem; step S6, constructing a contact environment model Env according to actual task conditions. stiff (z); Step S7, combining the environmental model, converting the actual task requirements into the form of convex constraints, and then converting the constraints and objective functions into expressions in the frequency domain, and finally using a standard convex optimization solver to solve the optimization problem and obtain the optimal controller under the closed-loop system.
Need to check novelty before this filing date? Find Prior Art