Multiple Targets Tracking Based On Intensity Filter

With the development of multi-target tracking technology, its application in the field of defense and civilian areas is increasingly widespread. The target can be modeled as a point target or extended target according to the different number of measurement generated by the same target. Sponsored by the National Natural Science Foundation of China,the dissertation mainly investigates the applications of the Intensity Filter and its special case-PHD filter in the field of point target tracking and extended target tracking. The main contributions of the dissertation are as follows:1.For unknown newly born targets, a novel passive multi-sensor target tracking algorithm based on Intensity Filter(iFilter) is proposed. All possible newly born targets are got by associating the continuous time of date. The states of targets are estimated and corrected by measurements.2.The implementation of Sequential Monte Carlo(SMC) for Marked Multitarget Intensity Filter(MMIF) is given. The simulations show that the algorithm is effective in point target tracking.3.The algorithm based on estimating the number of measurements for extended targets is proposed to solve the problem of unknown newly born targets. The number of measurements for each possible target is calculated and the targets owning more than one measurements are retained.4.The Iterative-Mapping PHD filter is proposed to solve the problem of measurement partition in extended target tracking. If the shape of extended targets is not considered, the extended targets can be tracked using the filter for the point targets. On the basis of Iterative-Mapping PHD filter, the algorithm based on linear fitting method is proposed to track the extended targets with unknown newly born targets.