Research On Multi-model Tracking Algorithm Of Hypersonic Glide Vehicle In Near Space

With the development of high speed aircraft in nearby space,its defense interception has attracted the attention of scholars all over the world.It is very difficult to track and intercept nearby space vehicles due to their strong maneuverability and super high speed.To solve this problem,this paper considers the target maneuver detection and combines the multi-model tracking algorithm to improve the tracking performance of strong maneuvering targets.In this paper,the target tracking principle and filtering method are first presented.The principles and characteristics of several linear and nonlinear filters are given.The common tracking models are also introduced in detail firstly.Secondly,the dynamic tracking model of the target is established with considering to the characteristics of the gliding phase of a high-speed and strong maneuvering target in nearby space fully.The aerodynamic characteristics and maneuvering state of the target are expanded into the state of the target tracking.And the space-based infrared detection model is adopted for positioning an d tracking.Effect of tracking of the tracking model is better than that of the kinematics model.The multi-model tracking method has a good tracking effect in the near space vehicle According to discussing the interactive multi-model algorithm and variable structure multi-model algorithm,this paper summarizes the characteristics of each method,and analyzes the tracking performance and limitations of each method through MATLAB simulation.Since the maneuvering time and the maneuvering form of the target are unknown,a multi-model tracking algorithm based on maneuvering detection is introduced.In the process of target tracking,the filtering residual error shows the tracking error very well.The square detection method is used to judge the target that is in the state of strong maneuvering,weak maneuveri ng or non-maneuvering,and then the multi-model algorithm model transfer and model selection are guided.In the aspect of model selection,variable structure multi-model method based on maneuvering detection is established.On the hand,Different maneuvering states correspond to different updating methods of tracking model set.On the other hand,the state transition probability between models is changed from the first-order Markov chain hypothesis to the hidden Markov model hypothesis.Based on the change of target maneuvering state,the adaptive updating of the transition probability between models is carried out,which makes up the actual error caused by the fixed transfer probability between models.The simulation results show that the tracking accuracy is improved,the model switching is faster and the overall tracking performance is better.In this paper,a dynamic modeling method for the tracking of hypersonic glide vehicle in nearby space is presented.At the same time,combined with multi-model filtering tracking,two multi-model tracking algorithms GIMM and HIMM based on maneuvering detection are presented under the premise of high speed and strong maneuvering,which improves the tracking performance.