Antenna Placement Optimization for Distributed MIMO Radar Based on a Reinforcement Learning Algorithm

Area of Science:

• Radar Systems Engineering
• Signal Processing
• Artificial Intelligence in Defense

Background:

• Distributed multiple-input multiple-output (MIMO) radar systems are crucial for target detection and tracking.
• Optimizing antenna placement is essential for maximizing system coverage, especially with dynamic targets.
• Traditional methods struggle with the complexities of varying target heading angles.

Purpose of the Study:

• To address the antenna placement optimization problem in distributed MIMO radar systems for multiple target heading angles.
• To develop an improved method for calculating system coverage area considering changing target headings.
• To leverage reinforcement learning for effective antenna placement solutions.

Main Methods:

• Mathematical modeling of antenna placement as a sequential decision problem.
• Implementation of a reinforcement learning agent utilizing the long short-term memory (LSTM)-based proximal policy optimization (PPO) algorithm.
• Development of an improved method for calculating radar system coverage area.

Main Results:

• The proposed reinforcement learning approach effectively solves the antenna placement optimization problem.
• The method demonstrates a significant enhancement in the achievable coverage area of the distributed MIMO radar system.
• Experimental findings validate the efficacy of the LSTM-PPO based agent.

Conclusions:

• The developed method provides a novel and effective solution for antenna placement optimization in distributed MIMO radar.
• The approach offers a valuable reference for future research and development in radar system design.
• Enhanced coverage area achieved through intelligent antenna placement contributes to improved radar performance.