Incomplete Measurement Of The Target Tracking Filter Design Under

Target tracking which could estimate information about target's location and track target in a class of hybrid coordinates is one of the hot research in and abroad China at present. It is widely used in the area of military, medical, navigation, security monitoring and so on.The filtering methods frequently used in target tracking based on hybrid coordinate are DCMKF and UKF. During the practical engineering, as the existence of variables of disturbances, the actual measurement data is lost to some extent inevitably which is called incomplete measurement. This will affect the precision and stability of the filter directly and even causes the lost of the tracking target. At present, domestic and foreign scholars have analyzed the statistical characteristics of standard Kalman filter theoretically under the situation of incomplete measurement. Based on the idea of incomplete measurement, this thesis designs two kings of filters called DCMKF and UKF under the situation of incomplete measurement and analyzes their statistical characteristics so that providing the theoretical basis of the filter design.In the first placement, in accordance with the problem of target tracking, we discuss the characteristics of nonlinear filter algorithms, and then we show the advantage and weak point of each filter through theoretical analysis and simulation verification and point out the reasons why DCMKF and UKF are proper methods in target tracking. Next, we extend the two kings of filters into the situation of incomplete measurement with the methods of residual detection and filter switch.In accordance with the two kinds of filters designed in the thesis, we bring about the analysis of the filter under statistical meanings. As to DCMKF under incomplete measurement, the thesis analysis and testifies the relationship between detection probability and estimation error variance under statistical meanings, pointing out that when detection probability is above a certain critical value, the filter is stable and meanwhile giving the upper bound of the critical value. When coming to the UKF under incomplete measurement, the thesis proves the existence of critical detection probability by analyzing the sufficient condition of stability under complete measurement and combining theoretical analysis and simulation verification, which is of some reference significance for engineering design.