Research On The Methods Of Track Initiation And Maneuvering Target Tracking

Nowadays, with the development of technology, the circumstance of modern war has changed a lot. The former methods were not able to meet the actual combat's need, especially the emergences of Network war and the development of information fusion. The circumstance of war is becoming more and more complicated, and there are more and more new weapons come forth, too. All above bring tracks management a big challenge. For these reasons, in this paper we mainly do some researches on track initiation,track fusion and target tracking under background of the project between a certain department of China Aerospace Science&Industy Company and us, and we also have done some forward-looking research on high maneuver targets tracking.Firstly, this paper describes the mostly track initiation methods for engineering usage. For the sequence track initiation algorithms can't be used under complex circumstance, and batch processing methods'computation burden is too large to carry out, a new algorithm based on"One Step Delay"model is presented, which adopts the adapted Multiple Hypothesis algorithm for data association, in order to be suit for the complex circumstance, multi-stage filtering is added in. Simulation results show that the algorithm which is used in this paper is easy for implementation, the ratio of correct track initiation is high, and the foreground of this algorithm's engineering application is better too.Secondly, a NFE model is presented and an algorithm for track-to-track fusion based on this model is proposed to solve the problem of the bad performance of common method for track fusion under complicated disturbing environment and radar netting dynamically. This algorithm, which applied the self-learning, self-reasoning ability and robustness of fuzzy neural network and expert system, can adapt to the changes of the environment and find the optimal algorithm. Simulation testified the effectiveness of this algorithm.Thirdly, the Interacting Multiple Model is analyzed for maneuvering targets tracking. More and more Models are added to the IMM algorithm for a better tracking precision, at the same time, this brings huge computation burden. For this reason, we mainly studied the Adaptive Grid Interacting Multiple Model and...