Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Multi-Robot Pursuit in Parameterized Formation via Imitation Learning

Jinyong Chen, Rui Zhou, Zhaozong Wang, Yunjie Zhang, Guibin Sun

TL;DR

A parameterized formation controller that allows defending robots to adapt their formation shape using five adjustable parameters is proposed and an imitation-learning based approach integrated with model predictive control to optimize these shape parameters is developed.

Abstract

This paper studies the problem of multi-robot pursuit of how to coordinate a group of defending robots to capture a faster attacker before it enters a protected area. Such operation for defending robots is challenging due to the unknown avoidance strategy and higher speed of the attacker, coupled with the limited communication capabilities of defenders. To solve this problem, we propose a parameterized formation controller that allows defending robots to adapt their formation shape using five adjustable parameters. Moreover, we develop an imitation-learning based approach integrated with model predictive control to optimize these shape parameters. We make full use of these two techniques to enhance the capture capabilities of defending robots through ongoing training. Both simulation and experiment are provided to verify the effectiveness and robustness of our proposed controller. Simulation results show that defending robots can rapidly learn an effective strategy for capturing the attacker, and moreover the learned strategy remains effective across varying numbers of defenders. Experiment results on real robot platforms further validated these findings.

Multi-Robot Pursuit in Parameterized Formation via Imitation Learning

TL;DR

A parameterized formation controller that allows defending robots to adapt their formation shape using five adjustable parameters is proposed and an imitation-learning based approach integrated with model predictive control to optimize these shape parameters is developed.

Abstract

This paper studies the problem of multi-robot pursuit of how to coordinate a group of defending robots to capture a faster attacker before it enters a protected area. Such operation for defending robots is challenging due to the unknown avoidance strategy and higher speed of the attacker, coupled with the limited communication capabilities of defenders. To solve this problem, we propose a parameterized formation controller that allows defending robots to adapt their formation shape using five adjustable parameters. Moreover, we develop an imitation-learning based approach integrated with model predictive control to optimize these shape parameters. We make full use of these two techniques to enhance the capture capabilities of defending robots through ongoing training. Both simulation and experiment are provided to verify the effectiveness and robustness of our proposed controller. Simulation results show that defending robots can rapidly learn an effective strategy for capturing the attacker, and moreover the learned strategy remains effective across varying numbers of defenders. Experiment results on real robot platforms further validated these findings.

Paper Structure

This paper contains 17 sections, 1 theorem, 26 equations, 13 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Problem Formulation
  3. Methodology
  4. Multi-robot Pursuit in Parameterized Formation
  5. Distributed Negotiation on Shape Parameter
  6. Imitation Learning for Shape Parameter
  7. Expert policy
  8. Imitation learning
  9. Simulation Results
  10. Simulation Setting and Learning Process
  11. Comparison Simulation
  12. Generalization Simulation
  13. Experiments
  14. Experiment Setting
  15. Shape Manipulation Experiment
  16. ...and 2 more sections

Key Result

Theorem 1

If the condition convergence_condition is satisfied, the negotiation strategy in Equ_negotiation_para can guarantee that the estimated shape parameters of all defenders can converge to the optimal value $\theta^*$, that is,

Figures (13)

  • Figure 1: The graphical illustration of the problem setting.
  • Figure 2: Parameterized formations representation.
  • Figure 3: The distributed implementation workflow for defenders.
  • Figure 4: Algorithm flow for updating the shape parameter.
  • Figure 5: Graphical explanation of cost function design. (a) The capture angle of a single. (b) Schematic diagram of capture angle, protected angle, and alignment angle.
  • ...and 8 more figures

Theorems & Definitions (1)

  • Theorem 1: convergence of policy negotiation