Greedy Satellite Scheduling Method with Minimal Switching Based on Interacting Multiple Model

This paper investigates the tracking of highly maneuverable targets during flight and the corresponding satellite scheduling problem in a space-based observation system. Based on dual-satellite measurements, a nonlinear observation equation was formulated. The J2 perturbation model and the current statistical model were utilized within an Interacting Multiple Model filtering framework to achieve adaptive estimation of the target states of the boost-glide vehicles. Building upon this framework, a greedy satellite scheduling algorithm based on IMM model probabilities is proposed. This method dynamically selects the optimal measurement set within a given prediction window to maximize observation performance. The proposed strategy is compared against rolling-horizon scheduling and fast-slow timescale scheduling approaches. Simulation results demonstrate that the proposed method effectively adjusts model weights in response to target maneuvers, enhancing adaptability during highly maneuverable phases. Meanwhile, it reduces the number of satellite switches while maintaining estimation accuracy, significantly improving scheduling efficiency and tracking continuity.