Visual-Based SLAM Configurations for Cooperative Multi-UAV Systems with a Lead Agent: An Observability-Based Approach

Area of Science:

• Robotics
• Computer Vision
• Control Systems

Background:

• Cooperative Simultaneous Localization and Mapping (SLAM) is crucial for multi-unmanned aerial vehicle (UAV) systems.
• Estimating the states of individual UAVs and a dynamic lead agent is a fundamental challenge.

Purpose of the Study:

• To address cooperative visual-based SLAM for multi-UAV systems with a lead agent.
• To develop robust state estimation for aerial robots and their lead agent.

Main Methods:

• Investigated three system configurations using nonlinear observability analysis.
• Proposed a high-level control scheme for formation control relative to the lead agent.

Main Results:

• Theoretical results and extensive computer simulations validated the proposed approach.
• Demonstrated robust performance in GPS-challenging environments.

Conclusions:

• The cooperative visual-SLAM method provides accurate position estimation for UAVs and lead agents.
• The approach is effective in various conditions, enhancing multi-UAV system capabilities.