基于ET-PHD的自适应联合跟踪与分类算法

针对新生目标强度先验未知的扩展目标（Extended target，ET）联合跟踪与分类（Joint tracking and classification，JTC）问题，提出一种基于扩展目标概率假设密度（Extended target-probability hypothesis density，ET-PHD）滤波器的自适应联合跟踪与分类算法，并给出其高斯混合实现方法.算法利用量测信息生成新生目标强度，在滤波预测阶段对存活目标和新生目标分别按照其类别进行传播，再引入属性量测信息，用位置和属性的联合量测似然函数代替单目标位置似然函数，对预测后所有目标强度进行联合更新，之后按照类别进行高斯项的删减与合并，提取相应类别目标的状态集.仿真结果表明，提出的自适应算法改进了概率假设密度滤波器在扩展目标跟踪中的性能.

Adaptive Joint Tracking and Classification Algorithm Using ET-PHD Filter

For joint tracking and classification (JTC) of extended targets and unknown target birth intensity, an adaptive algorithm based on the extended target probability hypothesis density (ET-PHD) filter is proposed with the Gaussian mixture implementation. The main idea is to approximate the birth intensity by using received measurements, and in the prediction stage the persistent and the newborn targets are distinguished to propagate according to their classes. Then the target classification information is integrated into the update stage and the combined measurement likelihood is used to substitute the single target position likelihood. A joint update is implemented by all predicted posterior intensities and then the Gaussian mixture components are pruned and merged according to their classes to characterize corresponding target state sets. Simulation results show that the adaptive algorithm can improve the performance of probability hypothesis density filter in the extended target tracking.