Fast Distributed Multiple-Model Nonlinearity Estimation for Tracking the Non-Cooperative Highly Maneuvering Target

The newly developed near-space vehicle has the characteristics of high speed and strong maneuverability, being able to perform vertical skips a wide range lateral maneuvers. Tracking this kind target with ground-based radars is difficult because limited detection caused by curvature Earth. Compared radars, satellite tracking platforms equipped Synthetic Aperture Radars (SARs) have range, can keep targets in custody, making them promising approach vehicles continuously. However, may not work well, due unknown maneuvers non-cooperative target, computing power satellites. To enhance stability accuracy, lower computational burden, we proposed Fast Distributed Multiple-Model (FDMM) nonlinearity estimation algorithm for satellites, which adopts novel distributed multiple-model fusion framework. This first requires each local filtering based on its own single model, corresponding factor derived Wasserstein distance solved estimate; then, after diffusing estimates, performs received minimum weighted Kullback–Leibler divergence; finally, updates state according consensus protocol. Two simulation experiments revealed that FDMM outperformed other four algorithms: consensus-based UKF; improved STUKF; strong-tracking adaptive CKF; interactive had precision low complexity, showing effectiveness satellites target.