Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Active Admittance Control with Iterative Learning for General-Purpose Contact-Rich Manipulation

Bo Zhou, Yuyao Sun, Wenbo Liu, Ruixuan Jiao, Fang Fang, Shihua Li

TL;DR

A new combined framework is proposed which introduces iterative learning control (ILC) strategy into the conventional admittance control and uses a new admittance parameter tuning approach, which endows the algorithm with automatic parameters tuning ability.

Abstract

Force interaction is inevitable when robots face multiple operation scenarios. How to make the robot competent in force control for generalized operations such as multi-tasks still remains a challenging problem. Aiming at the reproducibility of interaction tasks and the lack of a generalized force control framework for multi-task scenarios, this paper proposes a novel hybrid control framework based on active admittance control with iterative learning parameters-tunning mechanism. The method adopts admittance control as the underlying algorithm to ensure flexibility, and iterative learning as the high-level algorithm to regulate the parameters of the admittance model. The whole algorithm has flexibility and learning ability, which is capable of achieving the goal of excellent versatility. Four representative interactive robot manipulation tasks are chosen to investigate the consistency and generalisability of the proposed method. Experiments are designed to verify the effectiveness of the whole framework, and an average of 98.21% and 91.52% improvement of RMSE is obtained relative to the traditional admittance control as well as the model-free adaptive control, respectively.

Active Admittance Control with Iterative Learning for General-Purpose Contact-Rich Manipulation

TL;DR

A new combined framework is proposed which introduces iterative learning control (ILC) strategy into the conventional admittance control and uses a new admittance parameter tuning approach, which endows the algorithm with automatic parameters tuning ability.

Abstract

Force interaction is inevitable when robots face multiple operation scenarios. How to make the robot competent in force control for generalized operations such as multi-tasks still remains a challenging problem. Aiming at the reproducibility of interaction tasks and the lack of a generalized force control framework for multi-task scenarios, this paper proposes a novel hybrid control framework based on active admittance control with iterative learning parameters-tunning mechanism. The method adopts admittance control as the underlying algorithm to ensure flexibility, and iterative learning as the high-level algorithm to regulate the parameters of the admittance model. The whole algorithm has flexibility and learning ability, which is capable of achieving the goal of excellent versatility. Four representative interactive robot manipulation tasks are chosen to investigate the consistency and generalisability of the proposed method. Experiments are designed to verify the effectiveness of the whole framework, and an average of 98.21% and 91.52% improvement of RMSE is obtained relative to the traditional admittance control as well as the model-free adaptive control, respectively.
Paper Structure (15 sections, 1 theorem, 22 equations, 9 figures, 2 tables)

This paper contains 15 sections, 1 theorem, 22 equations, 9 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Multi-task Manipulation
  4. Interactive Force Control Methods
  5. Iterative Learning Methods
  6. Problem Formulation
  7. Platform and Operations
  8. Manipulation System Design
  9. Modeling of admittance towards parameter tuning
  10. ILC-MBK Algorithm
  11. Convergence Analysis
  12. Results of Experiments
  13. Experiments Preparation
  14. Experiment Results
  15. Conclusion

Key Result

Theorem 1

Considering the system (eq13) which satisfy the Assumption 1 and 2, similarly to the above matrix $G$, the matrix can be obtained: The infinity norm of $G_m$ matrix satisfy:

Figures (9)

  • Figure 1: Contact-rich movement primitives categorized into five classes, each of these contains multiple operations or multiple operation objects of the same class
  • Figure 2: Target switch box with three typical switches, auto reset button, second gear knob and emergency button
  • Figure 3: Hardware platform, contains Universal Robots UR5 robotic arm, OptoForce 6-Axis Force/Torque Sensor and a multi-functional end-effector
  • Figure 4: Filtered force-position relation, respectively pressing auto reset button, twisting second gear knob, pressing emergency button and resetting emergency button
  • Figure 5: ILC-MBK procedure during iterations, $u_k$ means the parameter input of admittance control, reference force position gives the reference to admittance control and the update law. The system output $y_k$ can be gotten after the interactive tasks, then the update law makes the $u_k$ go to the new generation $u_{k+1}$
  • ...and 4 more figures

Theorems & Definitions (3)

  • Remark 1
  • Theorem 1
  • Proof 1