基于支撑函数的不规则形态扩展目标建模和估计研究

项目名称： 基于支撑函数的不规则形态扩展目标建模和估计研究

项目编号： No.U1504619

项目类型： 联合基金项目

立项/批准年度： 2016

项目学科： 自动化技术、计算机技术

项目作者： 孙力帆

作者单位： 河南科技大学

项目金额： 27万元

中文摘要： 现代高精度传感器不但可以提供精确的目标运动状态而且还能获取目标散射中心沿视线方向上的高分辨率距离像，于是观测对象通常被建模成具有一定形态的扩展目标。传统上基于点目标假设所取得的建模与状态估计研究成果已不再适用于现今许多的实际应用场景，迫切需要对扩展目标的这一领域内的诸多理论和技术难点进行深入研究。但目前对不规则形态扩展目标，缺乏较为统一且精确的数学模型以及相对简洁而高效的估计算法。本项目针对不规则形态扩展目标的建模和估计及其相应的联合跟踪与分类进行以下研究：1）将不规则形态扩展目标描述成多个简单目标闵可夫斯基和的有机组合，并基于支撑函建立起通用模型以精确描述目标运动状态及其扩展形态；2）提出机动扩展目标混合系统模型并研究相应的多模型估计器；3）基于支撑函数在统一框架内研究解决距离像量测下的扩展目标联合跟踪与分类问题。针对上述研究问题所取得的进展和突破，对目标跟踪系统性能的提升有着重大意义。

中文关键词： 不规则扩展目标；支撑函数；目标距离像；闵可夫斯基和；联合跟踪与分类

英文摘要： Modern High resolution sensors can provide not only kinematic measurements but also one or more dimensions of the range extent measurements of an observed object along the line of sight. In this case, the object should be considered as an extended one with object extension. Thus, conventional point target modeling and state estimation approaches are no longer suitable for many current tracking scenarios, and further research on theory and technology of extended object tracking has become a pressing need. However, comparatively accurate unified models and efficient estimation approaches are still lacking for irregular extended objects. This project attempts to research on modeling and estimation of extended objects and its application to joint tracking and classification, which are summarized as follows: 1) An irregular extended object is represented as the minkowski addition of multiple simple objects, then a generalized model based on support functions is constructed to accurately describe the kinematic state and extension of the object. 2）Specially for tracking of maneuvering extended objects, a generalized model is proposed and the corresponding multiple-model-based estimator is then studied. 3) An unified framework based on support functions is further researched to solve joint tracking and classification of extended objects using the measurements of target range extent. Research process and breakthrough on the above topics could have great significance for improving the performance of target tracking.

英文关键词： Irrgular extended object;Support function;Target range extent;Minkowski addition;Joint tracking and classification