一种基于HMM的多传感器多目标快速跟踪算法  

A New Efficient Algorithm for Multisensor Multitarget Detection and Tracking Based on HMM

作　　者：李士民[1] 郭立[1] 朱嘉[1] 

机构地区：[1]中国科学技术大学电子科学和技术系,安徽合肥230026

出　　处：《系统工程与电子技术》2002年第5期84-87,共4页Systems Engineering and Electronics

基　　金：安徽省自然科学基金资助课题 ( 0 0 0 430 4)

摘　　要：在分布式多传感器系统中 ,一种基于隐Morkov模型的检测与跟踪算法得到了应用。该算法的优点在于适用于机动目标模型 ,且对目标的观测噪声未加任何限制。但由于该模型中所采用的Viterbi算法计算量大 ,实时性较差 ,因此提出了一种自适应算法以减少Viterbi算法中的冗余状态 ,有效地克服了其计算量大的缺点。实验表明 ,在相同的多目标跟踪仿真环境下 ,自适应Viterbi算法可以有效提高计算速度 。An algorithm based on HMM model has been developed for multiple-target detection and tracking in a distributed sensor system. The advantage of this approach is its robustness to maneuvers and impulsive noise. Considering the drawback of Viterbi algorithm in the model, this paper proposes an adaptive Viterbi algorithm to reduce the computational complexity. The simulation results show that the model which uses the adaptive Viterbi algorithm yields nearly the same performance as the model which uses the Viterbi algorithm while involving a substantially smaller average computation requirement in a similar environment.

关 键 词：多传感器多目标快速跟踪算法 卡尔曼滤波 隐马尔可夫模型 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]