Multi-UAV-Borne Surveillance Radar Trajectory Planning Method Based on Imitation Learning.

Area of Science:

• Robotics and Control Systems
• Artificial Intelligence
• Sensor Networks

Background:

• Traditional multi-radar trajectory planning faces challenges with computational complexity and real-time performance in complex scenarios.
• Effective trajectory planning is crucial for optimizing multi-platform sensing capabilities.

Purpose of the Study:

• To develop an efficient and real-time trajectory planning method for multi-radar systems using imitation learning.
• To improve the coverage rate and decision-making speed in multi-platform sensing.

Main Methods:

• Proposed an imitation-learning-based trajectory planning method for multi-radar systems.
• Developed a trajectory policy neural network architecture incorporating multi-semantic information.
• Constructed training data with coverage rate as the optimization objective.
• Utilized an imitation-learning algorithm with an auxiliary target for neural network training.

Main Results:

• Achieved an average coverage rate of 93.95% in simulations.
• Improved single-step decision efficiency by a factor of 6.7 compared to heuristic-based methods.
• Demonstrated superior real-time performance and computational efficiency.

Conclusions:

• The proposed imitation-learning method effectively addresses the limitations of traditional trajectory planning in multi-radar systems.
• The method offers significant improvements in coverage rate and decision efficiency for complex multi-platform sensing.
• This approach provides a viable solution for real-time trajectory optimization in dynamic environments.