Adaptive Turn-Rate Estimation for Maneuvering Target Tracking via Auxiliary Estimators

To address the challenge of tracking maneuvering targets that exhibit coordinated turns and other complex motion patterns, the first contribution of this paper is the development of a lightweight model set within the Interactive Multiple Model (IMM) framework. Subsequently, in order to adaptively estimate the target’s turn rate to determine the parameters of the Coordinated-Turn (CT) model, this paper proposes three auxiliary estimators: (1) indirect least-squares estimation, (2) pseudo-measurement-based Kalman filtering, and (3) a neural estimator based on bidirectional long short-term memory (Bi-LSTM) and self-attention mechanisms. By feeding the estimated turn rate back into the CT model, the proposed method alleviates model-target mismatch and enhances tracking accuracy. Simulation results demonstrate that the above methods can effectively accomplish turn-rate estimation and maneuvering-target tracking, offering stable performance and strong practical applicability. The neural estimator achieves the highest precision and robustness among all the proposed methods.