Compound weighted fusion evaluation and optimization of intelligent tracking algorithm in radar seeker

Area of Science:

• Artificial Intelligence
• Control Systems Engineering
• Radar Signal Processing

Background:

• Evaluating and optimizing radar seeker neural network (NN) tracking algorithms is crucial for enhancing guidance and control system performance.
• Existing methods may be influenced by hardware variations, complicating objective algorithm assessment.
• Dynamic tracking scenarios require adaptive evaluation metrics that account for performance fluctuations over time.

Purpose of the Study:

• To design a hierarchical weighted fusion evaluation and optimization scheme for radar seeker NN tracking algorithms.
• To mitigate hardware influence on algorithm performance evaluation.
• To develop a robust method for optimizing NN tracking algorithms based on comprehensive performance scoring.

Main Methods:

• A three-stage weighted fusion process is employed for comprehensive algorithm evaluation.
• Closed-loop performance indices are initially fused to eliminate hardware effects.
• Tracking indices are scored using a hybrid linear-nonlinear mechanism across different time periods and fused hierarchically.

Main Results:

• A parameter evaluation case set was designed and analyzed using the proposed fusion scheme.
• The case yielding the highest comprehensive score was identified.
• The radar seeker NN tracking algorithm was optimized based on the highest-scoring case, demonstrating improved effectiveness in experimental validation.

Conclusions:

• The proposed hierarchical weighted fusion scheme provides an effective method for evaluating and optimizing radar seeker NN tracking algorithms.
• The approach successfully isolates algorithm performance from hardware dependencies.
• Experimental results confirm the method's efficacy in enhancing tracking algorithm performance.