Design and Implementation of Experimental Platform for Multi-agent System

摘要

To achieve programmability, modularity, cost-effectiveness, and practical effectiveness in multi-agent body experimental processes, we propose an experimental control platform for multi-agent body robots based on ROS (Robot Operating System) and MATLAB/Simulink in this paper. Firstly, we discuss the experimental platform used in this study, which integrates NOKOV motion capture equipment and the Gazebo simulator. Information is transferred through the ROS-Simulink information bridge, allowing users to focus on designing control algorithms and interacting with experimental information during use. Furthermore, we provide various modeling and simulation examples under the designed experimental platform, including different types of multi-agent bodies such as UAV and UGV, different UGV models including the Turtle model in ROS and Mecanum wheel simulation model, and various trajectory tracking algorithms such as Pure Pursuit and self-designed tracking algorithms. Through simulating and controlling experimental examples, we have discovered that correct modeling of the multi-agent body under the experimental platform allows the model code built on the Simulink platform to be directly transferred to the ROS platform for implementation or cross-platform control. This significantly enhances the efficiency of developers in solving practical engineering problems and generates high utilization value and positive economic benefits.