Spatial-Temporal Reservoir Inference with Knowledge-based Enhancements (STRIKE)

The U.S. Army requires advanced radar signal processing methods for counterfire detection that enhance detection sensitivity for long-range threats, improve clutter rejection, and enable real-time adaptability while maintaining computational efficiency and total scan time. Traditional Space-Time Adaptive Processing (STAP) methods struggle with complex clutter environments and high-speed projectile tracking, while deep learning approaches require large labeled datasets and lack interpretability. To address this challenge, Orbitswarm proposes developing Spatial-Temporal Reservoir Inference with Knowledge-based Enhancements (STRIKE), a novel radar signal processing framework that integrates Reservoir Computing and programmable logic to improve target detection and tracking in dense clutter and jamming environments. Reservoir Computing enables rapid adaptation, reducing computational overhead and reliance on extensive training, while spatial-dynamics-based filtering provides knowledge-based enhancements for false target mitigation. In Phase I, Orbitswarm will demonstrate the feasibility of STRIKE through a study using laboratory simulation tools. The team will develop a test plan, conduct simulation-based evaluations using synthetic radar data, and benchmark performance against conventional approaches. Test results will be analyzed to assess the effectiveness of the proposed technique, and a report will describe its implementation and potential for real-world deployment. Phase I will establish the foundation for Phase II, where STRIKE will be further refined, validated against real-world radar datasets, and optimized for integration with a government systems integration lab in preparation for deployment on Army radar platforms.