Robust State Estimation Under Coupled Maneuver and Non-Gaussian Measurement Uncertainties: A Hierarchical Decoupling Adaptive Fusion Approach
Abstract
State estimation for maneuvering targets becomes challenging when maneuver-induced model mismatch and non-Gaussian measurement disturbances occur simultaneously. These two uncertainty sources are physically different but coupled in the innovation sequence: maneuver residuals may be misidentified as measurement outliers, while aggressive noise suppression may attenuate useful maneuver information. To address this problem, this article proposes a hierarchical decoupling adaptive fusion (HiDAF) method that treats the two uncertainties sequentially. Specifically, an adaptive strong tracking filter (ASTF), calibrated via a state cumulative error (SCE) criterion, first provides a maneuver-responsive estimate, which is then progressively refined through innovation-reliability weighting, residual-lag correction, and covariance-aware recombination. Theoretical analysis establishes steady-state optimality preservation, covariance dominance, and multilevel consistency for the proposed method. Simulation results under Rayleigh–Gaussian mixture measurement noise show that HiDAF reduces the position and velocity average root mean square errors (ARMSEs) by <inline-formula> <tex-math notation="LaTeX">$20.7 {\,}\%$ </tex-math></inline-formula> and <inline-formula> <tex-math notation="LaTeX">$32.1 {\,}\%$ </tex-math></inline-formula>, respectively, over the baseline strong tracking filter (STF). Relative to the robust improved robust student’s <inline-formula> <tex-math notation="LaTeX">$t$ </tex-math></inline-formula> maximum correntropy criterion IRSTMCC filter, HiDAF reduces the position and velocity ARMSEs by <inline-formula> <tex-math notation="LaTeX">$20.9 {\,}\%$ </tex-math></inline-formula> and <inline-formula> <tex-math notation="LaTeX">$26.2 {\,}\%$ </tex-math></inline-formula>, respectively, while reducing the per-step runtime by <inline-formula> <tex-math notation="LaTeX">$68.9 {\,}\%$ </tex-math></inline-formula>. These results demonstrate that HiDAF provides a favorable balance among maneuver responsiveness, robustness to heavy-tailed disturbances, and real-time feasibility.