Multi-agent cooperative target search.

Area of Science:

• Robotics and Automation
• Computer Vision
• Distributed Systems

Background:

• Cooperative search by unmanned aerial vehicles (UAVs) is crucial for monitoring mobile ground targets.
• Limited sensing and communication capabilities of UAVs pose significant challenges.
• Target detection performance is affected by UAV altitude and camera field of view.

Purpose of the Study:

• To develop a vision-based cooperative search strategy for multiple mobile ground targets using a group of UAVs.
• To address the limitations of sensing and communication in UAVs during search operations.
• To optimize collective coverage area and detection performance for enhanced target search.

Main Methods:

• Dividing the surveillance region into cells to create probability maps for each UAV.
• Proposing a distributed probability map updating model incorporating information fusion, neighbor sharing, and decay.
• Formulating the search problem as a multi-agent cooperative coverage control problem.
• Implementing a distributed cooperative control scheme with neighbor communication constraints.

Main Results:

• The proposed distributed probability map updating model effectively integrates measurement data and environmental changes.
• The cooperative control scheme optimizes coverage and detection performance for multiple UAVs.
• Simulations demonstrate the effectiveness of the developed algorithms in cooperative target search scenarios.

Conclusions:

• The developed distributed algorithms enable efficient vision-based cooperative search for mobile ground targets by UAVs.
• The approach effectively handles limited sensing and communication constraints.
• The strategy enhances overall search effectiveness and target detection probability.