A Novel Motion Model Classification Network and Error-Corrected Extended Kalman Filter for Maneuvering Target Tracking

In the domain of maneuvering target tracking, traditional algorithms often rely on multiple predefined mathematical models to approximate target dynamics. However, the inherent randomness and unpredictability of target maneuvers pose significant challenges to accurate motion modeling. To address these challenges, this paper proposes a novel tracking algorithm that enhances adaptability to frequent state transitions by identifying target motion patterns. Furthermore, tracking accuracy is improved by correcting the tracking error of the extended Kalman filter (EKF). Simulation results demonstrate that the proposed method achieves superior stability and localization accuracy compared to conventional approaches.