高超声速滑翔飞行器多模型鲁棒跟踪方法

A novel cubature Huber-based filter based on digraph switching variable structure multiple model（CHFDSVSMM）is proposed for hypersonic gliding vehicle（HGV），which has the characteristics of strong maneuverability，changeable maneuvering form and uncertain maneuvering timing. A set of maneuvering models including adaptive nonzero mean damping oscillation（ANMDO）model and current statistical model（CSM）is developed for the characteristics of jump gliding motion. Considering that the model set of the fixed structure multi-model algorithm has the disadvantages of competition and long operation time，a switching criterion based on digraph structure is designed to change the composition of model set adaptively and improve model matching. In order to solve the problem of glint noise in the detection of HGV by ground-based radar，the cubature Huber-based filtering method is used to estimate the state. Simulation results show that compared with the existing methods，the proposed algorithm has higher tracking accuracy and better robustness to glint noise.