Research On Extended Target Tracking Algorithm Based On Random Hypersurface Model

In traditional target tracking methods,the target is usually regarded as a point target,ignored its shape feature and only estimated the kinematical state.With the wide application of high resolution sensors,sensors can acquire more than one measurement from the target's surface at each sample period.At this point,most of traditional target tracking algorithms become unsuitable because they are difficult to accurately describe the complete state of targets.Extended target tracking appears under this opportunity,and it becomes a hot research direction in target tracking area.Using the random finite set filter,aiming at the problem of how to estimate the extension state in Bernoulli,CBMe MBer and LMB filter,the thesis study extended target tracking algorithm based on random hypersurface model.The main research contents and achievements of this thesis are as follows:1.In order to solve the problem that the extension state estimation is inaccurate in the process of single elliptical extended target tracking,an algorithm of single extended target tracking based on random hypersurface model and Bernoulli filter using SMC implementation is proposed.The algorithm combines random hypersurface model and Bernoulli filter,modeling the measurement source with random hypersurface model.By the distribution of measurement source,the algorithm can describe the size of target expansion,and it can also make fully use of measurement information estimating the extension of target.Experimental results show that the proposed algorithm has a better performance than the existing algorithm,in the meantime,it can improve accuracy of the extension state estimation,also the proposed algorithm has excellent realistic signification.2.In view of the ET-CBMe MBer filter cannot estimate the extension of targets,an algorithm of extended targets tracking based on ET-CBMe MBer and random hypersurface model using SMC implementation is proposed.The algorithm obtains the distribution of measurement source through the random scale factor of elliptical random hypersurface model,which reasonably expresses the absence of the target's prior information.The parameters describing extension state are integrated into the kinematical state,which avoids the processing of matrix.Experimental results show that the proposed algorithm improves the accuracy of extended targets tracking and estimates the extension state of targets accurately.Meanwhile,it has great values on practice.3.Aiming at solving the problem that the traditional extended targets tracking algorithm cannot form effective trajectory of each target,an extended targets tracking algorithm based on random hypersurface model and LMB filter using Gamma Gaussian Mixture implementation is proposed.The core of the proposed algorithm is to model the measurement source using random hypersurface model before the updating of state,and to reflect the expansion degree of the target by the distribution of the measurement source model.Experimental results show that the proposed algorithm improves the accuracy of kinematical state estimation and extension state estimation.Moreover,it has good prospects for engineering application and great practical values.