Variable impedance control method for robot contact force based on TD3 algorithm

As a common method of traditional force control, impedance control has been widely used in robot machining field. However, impedance control with fixed parameters cannot dynamically adapt to the manufacturing environment. This paper presents a variable impedance control method for contact forces based on the TD3 (Twin Delayed Deep Deterministic Policy Gradient) algorithm. The method builds on the original impedance model to obtain the optimal impedance parameters for each stage by stochastic exploration of the continuous action space, reducing the reliance on human experience and multiple attempts in the selection of traditional impedance parameters. In this paper, gantry robot is taken as the research object. The simulation results show that the variable impedance control method has faster adjustment time and more stable, and the average position tracking error is about 0.11cm and the force tracking error is about 0.01N.

• Publication year

  2022

• Venue

  ACM Cloud and Autonomic Computing Conference

• Publication date

  2022-11-25

• Fields of study

  Not labeled

• Identifiers
  DOI 10.1109/CAC57257.2022.10055172
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Deep Deterministic Policy Gradient with Reward Function Based on Fuzzy Logic for Robotic Peg-in-Hole Assembly Tasks

  2022cited by this paper
• Adaptive Variable Impedance Control with Fuzzy-PI Compound Controller for Robot Trimming System

  2022cited by this paper
• Learning Variable Impedance Control via Inverse Reinforcement Learning for Force-Related Tasks

  2021cited by this paper
• Position-Based Fractional-Order Impedance Control of a 2 DOF Serial Manipulator

  2021cited by this paper
• A Path Planning Method of Robotic Belt Grinding for Workpieces With Complex Surfaces

  2020cited by this paper
• Research on Design and Motion Control of 5-DOF Robot for Bending

  2018cited by this paper
• Braverman Readings in Machine Learning. Key Ideas from Inception to Current State: International Conference Commemorating the 40th Anniversary of Emmanuil Braverman's Decease, Boston, MA, USA, April 28-30, 2017, Invited Talks

  2018cited by this paper
• Active Impedance Control of Bioinspired Motion Robotic Manipulators: An Overview

  2018cited by this paper
• Addressing Function Approximation Error in Actor-Critic Methods

  2018cited by this paper
• From Reinforcement Learning to Deep Reinforcement Learning: An Overview

  2017cited by this paper
• Variable Admittance Control Based on Fuzzy Reinforcement Learning for Minimally Invasive Surgery Manipulator

  2017cited by this paper
• Deterministic Policy Gradient Algorithms

  2014cited by this paper
• Robust control of dynamically interacting systems

  1988cited by this paper

• No citing papers are available for this paper.