A Screened IMM-GRU Method for Low-Probability-of-Intercept Maneuvering Target Tracking

Area of Science:

• Engineering
• Computer Science
• Signal Processing

Background:

• Accurate maneuvering target tracking is challenging due to difficulties in prespecifying motion models and delayed model switching.
• Existing methods often struggle with real-time adaptation to target maneuvers, impacting tracking precision and radar resource management.

Purpose of the Study:

• To propose a joint model- and network-driven screened Interacting Multiple Model-Gated Recurrent Unit (IMM-GRU) for improved maneuvering target tracking.
• To enhance radar low-probability-of-intercept (LPI) operation by accurately identifying target motion modes and optimizing radar emissions.

Main Methods:

• A screened IMM-GRU method incorporating a novel filtering stage that uses a local reference trajectory for estimate validation.
• A difference-enhanced attention Bidirectional GRU (Bi-GRU) classifier for target motion mode identification during radar-silent intervals.
• A screened IMM-GRU predictor that selects motion model candidates based on multiple criteria including kinematic consistency and prediction uncertainty.

Main Results:

• The proposed filtering and prediction modules significantly improve tracking and extrapolation accuracy for maneuvering targets.
• The framework achieves the lowest average radar radiation ratio under the equal-accuracy LPI criterion across various MRMSE levels.
• The method uniquely satisfies the stringent 30 m MRMSE requirement in all tested scenarios.

Conclusions:

• The developed screened IMM-GRU method offers superior maneuvering target tracking performance compared to existing approaches.
• The framework effectively reduces unnecessary radar exposure, crucial for LPI operations.
• This approach demonstrates a significant advancement in adaptive tracking systems for complex target dynamics.