Research On Adaptive Tracking Algorithm For Maneuvering Target Of Near Space

Maneuvering targets of the near space,which have high speed,strong maneuverability and variable movement patterns traits,put forward more higher requirements for performance of tracking system.In order to solve the tracking problems brought by vehicles in near space,considered from the following aspects:1.In tracking system,due to the uncertainties of state model,the accuracy and stability of state estimation algorithm are reduced.For such problems,attenuation memory cubature Kalman filter is proposed,and the feasibility of proposed algorithm for achieving adaptive is analyzed.2.The ski-jump flight is a common motion pattern for vehicles in near space.For this pattern,it is combined with sinusoidal auto correlation hypothesis in Sine Wave model that the maneuvering acceleration is regarded as a non-zero mean stochastic process.An adaptive tracking model with correlation noise is proposed,and the effectiveness of proposed model is illustrated by simulations.3.The tracking error is large when turning targets with high speed and maneuverability are tracked by two-dimensional turn model,which is modeled under a constant rate.For these targets,the parameters in covariance matrix of three-dimensional variable turn model are modified,also,its effect of adaptive identification and tracking accuracy are validated in simulation cases.4.There are generally multi-phased rapidly changed maneuvers for complex moving targets.Single model algorithm is difficult to keep continuous tracking.For this case,the principles and differences of fixed structure and variable structure multiple model algorithms are analyzed,also,the improved performance brought by interacting multiple model and its necessity are proved by simulations.