Cooperative search algorithm for UAV swarm based on heterogeneous sensor fusion.

Area of Science:

• Robotics and Autonomous Systems
• Sensor Fusion
• Artificial Intelligence

Background:

• Single-sensor detection in complex environments lacks robustness.
• Unmanned Aerial Vehicle (UAV) swarms offer potential for enhanced search capabilities.
• Heterogeneous sensor fusion can improve detection accuracy and reliability.

Purpose of the Study:

• To develop a cooperative search algorithm for UAV swarms utilizing heterogeneous sensor fusion (HS-CS).
• To enhance the robustness and effectiveness of target detection in complex environments.
• To improve the fusion coverage and difficult target detection rates for UAV swarms.

Main Methods:

• Discretizing the mission area into a grid and constructing a four-state map model.
• Designing collaborative update and distributed fusion operators for accurate map updates.
• Employing a fast non-dominated sorting approach for dual optimization objectives (total and fusion coverage).
• Defining a multi-dimensional evaluation index with a four-stage adaptive function and stochastic exploration.

Main Results:

• Achieved an average fusion coverage rate of 97% in simulations with 50 targets.
• Reached an average difficult target detection rate of 97.1% for targets requiring fused detection.
• Demonstrated high performance over 50 independent repeated experiments.

Conclusions:

• The HS-CS algorithm significantly improves cooperative search capabilities for UAV swarms.
• Heterogeneous sensor fusion is effective in enhancing detection robustness in complex scenarios.
• The proposed algorithm shows strong potential for practical cooperative search applications.