IMMNet: Hybrid Fusion of Model-based and Data-driven Approaches for Maneuvering Target Tracking

arXiv cs.AI · 18 hours ago

Researchers proposed IMMNet, a hybrid algorithm combining the interacting multiple model framework with neural network components for tracking maneuvering targets in three-dimensional space. The method maintains Bayesian inference while adding learnable components to adapt to motion patterns and noise characteristics, and demonstrated superior performance compared to existing algorithms across multiple test scenarios. This approach enables real-time radar applications to better handle complex target movements while preserving interpretability and practical reliability.