基于随机矩阵的机动不规则扩展目标建模与估计研究

项目名称： 基于随机矩阵的机动不规则扩展目标建模与估计研究

项目编号： No.61203120

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 自动化学科

项目作者： 兰剑

作者单位： 西安交通大学

项目金额： 24万元

中文摘要： 现代传感器可对运动体多个测量点进行观测，而分辨率实际受限又使得测量点是部分可分辨的。这类运动体通常被建模为扩展目标。对不规则扩展目标机动过程中的运动状态及扩展形态进行精确估计已成为当前目标跟踪系统亟待解决的重要问题，应用潜力巨大。但目前对该问题的研究还不够深入，缺乏较为统一而精确的数学模型，以及相对简洁而高效的估计方法。本项目针对机动不规则扩展目标的建模与估计进行研究：将不规则目标表征为多个规则目标的空间组合，基于随机矩阵提出通用模型以精确描述目标状态和形态的时间演化过程及形态观测扭曲现象；继而就不规则目标整体机动过程建立混合系统模型；基于通用模型，研究单模型递推式贝叶斯状态及形态估计器；并针对混合系统设计变结构多模型状态及形态估计器，实现不规则扩展目标机动全过程的精确跟踪。针对上述研究问题取得的进展和突破，必将完善扩展目标建模与估计理论，并对提升信息融合及目标跟踪系统性能具有重大意义。

中文关键词： 扩展目标跟踪；状态估计；形态估计；联合跟踪与分类；性能评估

英文摘要： Multiple measurement points on a moving body can be observed by modern sensors. Due to limited sensor resolution, these points are also partially resolvable. Such a moving body is always modeled as an extended object. For current target tracking systems, accurately estimating the motion state and the extension of an irregular extended object in a maneuver process becomes an important problem which needs to be solved urgently. Related estimation technologies also have great application potential. However, this problem has not been studied thoroughly. Comparatively accurate unified models and efficient estimation approaches are still lacking. This project attempts to do researches on the modeling and the estimation of a maneuvering irregular extended object. Specifically, an irregular object is represented as a spatial combination of multiple regular objects first. Then a generalized model using random matrices is proposed to accurately describe the temporal evolution processes of the state and the extension, and also the observation distortion of the extension. The whole maneuver process of the irregular object is modeled by a hybrid system. Based on the generalized model, a single-model Bayesian state and extension estimator is studied. Then based on the hybrid system, a variable-structure multiple-model state a

英文关键词： Extended object tracking；State estimation；Extension estimation；Joint tracking and classification；Performance evaluation