Skip to main content

Distributed Robust Formation Control for Multi-Agent Systems - SICE 2024 Paper

· 4 min read
Guangze Yang
Guangze Yang
Master of Control System, Ibaraki University

Distributed Robust Time-Varying Formation Control for Multi-Agent Systems under Disturbances

Authors: Guang-Ze Yang¹ and Zi-Jiang Yang²

DR-TVFC Framework

Abstract​

This work considers the problem of time-varying formation tracking control of second-order multi-agent systems under disturbances. The DR-TVFC (Distributed Robust Time-Varying Formation Control) approach is proposed, including distributed finite-time estimators of the leader's states and sliding mode time-varying formation controllers.

:::info Publication Status This work has been accepted by the 2024 63rd Annual Conference of SICE. :::

Research Framework​

The framework demonstrates the integration of distributed estimation and robust control techniques for achieving time-varying formation control in multi-agent systems.

Source Code and Implementation​

:::tip GitHub Repository Source Code: https://github.com/youkoutaku/DR-TVFC

The complete implementation includes MATLAB simulation code, documentation, and visualization tools. :::

Program Structure​

Core Configuration​

  • config.m: System parameters and control parameters configuration

Signal Generation​

  • Leader_state.m: Generate reference signal (trajectory) and plot figures
  • Formation_shape.m: Generate time-varying formation shape and plot figures

Main Implementation​

  • main.m: Main simulation file containing:
    • Distributed finite-time estimator implementation
    • Prescribed time modification functions
    • Sliding mode formation controller
    • Complete system simulation

Visualization Programs (Plot/ directory)​

Main Figures​

  1. fig_motion.m: 3D plotting of agents' movement trajectories
  2. DR-TVFC_PPT.pdf: Presentation slides from the SICE 2024 conference
  3. fig_error_e.m: Estimation errors of the distributed estimator
  4. fig_states.m: State evolution of all agents
  5. fig_error_pro.m: Formation controller tracking errors
  6. fig_input.m: Control input signals
  7. Animation.m: Real-time animation of agents' movement (run after fig_states.m)

Additional Analysis Figures​

  1. fig_states_e.m: Estimator state evolution
  2. fig_vinput.m: Virtual input signals of the estimator
  3. fig_em.m: Modified formation tracking errors
  4. fig_xi.m: Prescribed time modification function visualization
  5. fig_sliding.m: Sliding variable evolution

Running the Simulation​

To execute the complete simulation and analysis:

build.m

This script will run the entire simulation pipeline and generate all visualization results.

Key Technical Features​

1. Distributed Finite-Time Estimation​

  • Sliding Mode Based: Robust estimation against disturbances
  • Finite-Time Convergence: Fast convergence to true leader states
  • Network Communication: Efficient information exchange protocol

2. Prescribed Time Modification​

  • Time-Varying Formation: Dynamic formation shape changes
  • Prescribed Convergence: Guaranteed convergence within specified time
  • Disturbance Rejection: Robust performance under external disturbances

3. Sliding Mode Control​

  • Robust Control: Inherent robustness to uncertainties and disturbances
  • Finite-Time Convergence: Fast system response
  • Chattering Reduction: Advanced techniques to minimize control chattering

Development History and Updates​

Recent Updates​

2024-08-01

  • Enhanced Animation.m program for improved presentation capabilities

2024-06-25

  • Globalized agent states to utilize matrix computation instead of iterative loops
  • Improved computational efficiency

Major Milestones​

2024-01-21

  • Implemented time-varying formation shape using SMFC algorithm

2023-11-26

  • Introduced distributed state estimation for consensus tracking control
  • Enhanced leader state estimation capabilities

2023-09-08

  • Integrated prescribed time modification function to formation tracking errors in SMFC

2023-07-30

  • Developed sliding mode formation controller (SMFC) for centralized leader-follower architecture

Ongoing Research​

2023-03-11

  • Collision avoidance for multi-agent systems (ongoing development)

2023-11-06

  • Neural network integration for disturbance handling (experimental phase)

Applications and Impact​

Robotics Applications​

  • Autonomous Vehicle Formation: Coordinated movement of autonomous vehicles
  • Drone Swarms: Formation flying and collaborative missions
  • Mobile Robot Teams: Warehouse automation and logistics

Industrial Applications​

  • Manufacturing Systems: Coordinated robotic assembly lines
  • Process Control: Multi-unit chemical process coordination
  • Smart Grid: Distributed energy system coordination

Research Significance​

  • Theoretical Advancement: Novel approach to distributed robust control
  • Practical Implementation: MATLAB-based tools for research and education
  • Community Contribution: Open-source implementation for reproducible research

Future Directions​

  1. Enhanced Collision Avoidance: Integration of sophisticated obstacle avoidance algorithms
  2. Machine Learning Integration: Neural network-based disturbance estimation and compensation
  3. Experimental Validation: Hardware implementation and real-world testing
  4. Scalability Studies: Performance analysis with larger agent populations

Author Information​

¹ Guang-Ze Yang
Department of Mechanical Systems Engineering, Ibaraki University, Hitachi, Japan
Tel: +81-294-38-5205
Email: 24nm499s@vc.ibaraki.ac.jp

² Zi-Jiang Yang
Department of Mechanical Systems Engineering, Ibaraki University, Hitachi, Japan
Tel: +81-294-38-5205
Email: shikoh.yoh.zijiang@vc.ibaraki.ac.jp