Visual Collaboration Leader-Follower UAV-Formation for Indoor Exploration

Area of Science:

• Robotics
• Computer Vision
• Aerospace Engineering

Background:

• Unmanned Aerial Vehicles (UAVs) in leader-follower formations require precise relative pose information.
• Traditional RF communication for pose data introduces latency and packet loss.
• Existing methods lack robustness in dynamic formation scenarios.

Purpose of the Study:

• To develop and validate a novel visual-based relative pose estimation system for UAVs.
• To eliminate reliance on RF communication for inter-UAV data exchange in formations.
• To enhance the robustness and efficiency of UAV formation control.

Main Methods:

• Utilizing passive markers in a RhOct configuration on each UAV.
• Implementing on-board visual detection and relative pose determination.
• Developing a reference path for follower UAVs to track the leader.

Main Results:

• Achieved a mean position detection error below 0.0031% within the operational workspace.
• Demonstrated successful leader tracking and inter-member distance maintenance.
• Validated the system's tolerance to communication delays up to 1.25 seconds.

Conclusions:

• The visual marker system effectively replaces RF communication for relative pose estimation in UAV formations.
• This approach enhances formation control accuracy and resilience to communication latency.
• The method offers a promising solution for robust and efficient collaborative UAV operations.