Answers are grounded in this paper and cite their sources.
Navigation and maneuvering target tracking are fundamental to autonomous systems such as unmanned aerial vehicles, unmanned ground vehicles, intelligent robots, aircraft guidance systems, and radar surveillance systems, with the core task of achieving accurate and reliable estimation of key states including position, velocity, acceleration, and attitude. Under high-speed maneuvers, complex environments, and multisensor fusion conditions, nonlinearities, model uncertainties, and time-varying environmental disturbances may degrade the estimation accuracy of the conventional Extended Kalman Filter (EKF) and even cause filter divergence. Although the Strong Tracking Filter (STF) exhibits strong adaptability to model mismatch, it mainly compensates for prediction errors by adjusting the predicted error covariance matrix, without directly addressing uncertainties in the state-transition model. Consequently, its estimation performance remains limited in complex dynamic environments. To address these issues, this paper uses weighted least squares (WLS) to identify an equivalent multiplicative state-transition uncertainty factor and directly compensate the prediction model. The compensation targets model mismatch that can be approximated by scalar scaling of the nominal state-transition term. Furthermore, by integrating state-model information with the statistical characteristics of uncertain factors, an information-augmented state prediction method and a corresponding error covariance calculation method are proposed. By directly compensating the state prediction model, the proposed method aims to improve the consistency and accuracy of state prediction and covariance propagation under maneuvering conditions and external disturbances. On this basis, a novel strong tracking filter is developed within the EKF–STF framework by integrating WLS-based identification of multiplicative state-transition uncertainty, state prediction compensation, and covariance-associated information enhancement. The conventional one-way inflation mechanism of the STF is extended to a bidirectional adjustment mechanism capable of both inflation and contraction, enabling the filter to adapt to variations in model uncertainties. This allows the filter to balance tracking adaptability under maneuvering conditions with estimation accuracy under steady-state conditions. Finally, a conditional covariance analysis characterizes the ESTF-1–STF comparison under the stated assumptions. Numerical simulations show improved estimation accuracy for ESTF-1 relative to the selected STF and EKF baselines in the tested scenarios, with ESTF-0 included as an oracle reference.
No takes yet. Share an insight, caveat, or question.
Li et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: