Distributed Adaptive Sliding Innovation Fixed-Lag Smoother for Maneuvering Target Tracking with Impulsive Measurement Outliers

Target maneuvers and impulsive measurement outliers can both produce large innovations, posing a challenge to accurate space target tracking. To address their simultaneous occurrence, this study proposes a distributed adaptive sliding innovation fixed-lag smoother based on the consensus-based cubature sliding innovation filter (CCSIF). The proposed method retains cubature-based moment propagation and consensus fusion while replacing saturation-based gain modulation with Gaussian-kernel innovation weighting. Spatial synchrony across observation satellites and temporal persistence within a reference-innovation window are constructed to characterize innovation anomalies. These complementary features drive a fuzzy inference system (FIS) that adaptively adjusts the kernel width to balance outlier suppression and maneuver responsiveness. Subsequent measurements within the window guide the update of a past state, yielding a fixed-lag smoothing scheme. Simulations involving four low-Earth-orbit satellites tracking a maneuvering high-altitude space target under impulsive measurement contamination demonstrate substantially improved position estimation. With a fixed lag of 10 steps, the proposed smoother achieves time-averaged position and velocity root-mean-square errors (RMSEs) of 13.14 m and 0.1165 m/s over the final 200 s, respectively.