Maneuvering Target Tracking Research And Application Of Multi-model Filtering Algorithm

With the rapid development of modern weapons, the types of the targets' maneuvers are complicated,so tracking of maneuvering target is becoming more and more difficult.Target tracking is the technology that using the sensor to measure target information and estimating the target motion state accurately. Target tracking technology is widely used, both in military and civil fields. It has a very wide range of applications. Sensor technology and information technology have developed rapidly in recent years, especially in the continuous improvement of the modern control theory. A variety of new theories have been applied in the field of target tracking, so that target tracking technology has made considerable progress.This paper mainly studies target tracking filter when the target move as 'snake'. Taking into account of that the target maneuvering will change the motion model, using a single model to track filtering will produce large errors. Through comprehensive consideration,interactive multiple models filtering algorithm for a 'snake' maneuver target tracking filter is the most cost-efficient way. An improvement of the interactive multiple models filtering algorithm will be proposed in this paper to improve the tracking accuracy.Firstly, the mathematical modeling of maneuvering targets is introduced in related fileds. According to the complex forms of maneuvering targets, some simple models for target tracking movement are presented in this paper, such as, constant velocity model,constant acceleration model and model of turning at a constant speed etc. Then,a brief introduction to the filtering algorithm is conducted, from the most basic of least squares estimation to the Kalman filter and the extended form of Kalman filter algorithm. This knowledge had done the groundwork for the choice of filtering method. Finally, simulation and analysis for the maneuver target tracking based on interactive multiple models filtering algorithm have been done in this paper, using matlab. Aiming at the defects of the interacting multiple models filtering algorithm, the interactive multiple models filtering algorithm is improved by combining fuzzy control theory to it.In this paper,improvement of the interacting multiple models filtering algorithm is mainly based on the transfer problem of model probability. Through the integration of the fuzzy control theory with IMM algorithm, based on the defect of interactive multiple models filter algorithm and the experience of the tracking filter, an updating module of model probability based on fuzzy logic has been designed. It has increased the transfer rate of model probability and the tracking precision of the interactive multiple models algorithm.