Direct Position Tracking for Non-Circular Signals: A Chunked Adaptive Approach Under Measurement Mismatch

Traditional passive position tracking methods often assume stationary transmitters during brief periods of observation. However, this assumption does not hold in dynamic environments, leading to performance degradation, especially when tracking fast-moving emitters or in scenarios with varying signal noise levels across different receivers. In this paper, we propose a robust approach for tracking non-circular (NC) signals using distributed passive arrays, where single-snapshot measurements are leveraged for position estimation. Our method introduces a tracking system model that incorporates state expansion and signal fusion. To resolve ambiguities in the measurement-state relationship—caused by unknown source signals—we apply a weighted least squares (WLS) method for signal estimation. Finally, we derive an adaptive noise covariance update formula by leveraging the fusion properties of received signals from the distributed arrays, addressing challenges introduced by measurement mismatches due to source signal estimation errors. The proposed method is implemented within the unscented Kalman filter (UKF) framework. Extensive simulations and analysis of the posterior Cramér-Rao Lower Bound (PCRLB) demonstrate that our approach significantly improves tracking accuracy and reliability in complex environments.