Cooperative Monocular-Based SLAM for Multi-UAV Systems in GPS-Denied Environments

Area of Science:

• Robotics
• Computer Vision
• Navigation Systems

Background:

• Simultaneous Localization and Mapping (SLAM) is crucial for autonomous navigation.
• Multi-unmanned aerial vehicle (UAV) systems require robust localization, especially in GPS-denied environments.
• Monocular vision-based SLAM presents challenges in scale ambiguity and observability.

Purpose of the Study:

• To develop and validate a cooperative monocular-based SLAM approach for multi-UAV systems.
• To enhance the observability of multi-UAV systems using inter-vehicle visual measurements.
• To demonstrate improved localization accuracy in GPS-denied scenarios.

Main Methods:

• A cooperative SLAM framework utilizing monocular cameras on UAVs.
• Integration of relative distance measurements between UAVs, derived from visual data.
• Nonlinear observability analysis for theoretical validation.
• Extensive computer simulations to evaluate performance.

Main Results:

• The proposed cooperative approach significantly improves system observability compared to other visual SLAM configurations.
• The system demonstrates robust position and orientation estimation for UAVs in formation.
• Visual information and relative distance measurements enhance localization accuracy.

Conclusions:

• Cooperative monocular-based SLAM is a viable solution for multi-UAV navigation in GPS-denied environments.
• The integration of relative visual measurements enhances the observability and performance of the system.
• The approach provides accurate state estimation for coordinated aerial vehicle operations.