多假设跟踪算法研究及其应用

Reseach on Multiple Hypothesis Tracking Algorithm and Its Application

MHT(多假设跟踪)是一种基于多个扫描周期量测,进行数据互联的技术,理论上是解决数据关联的最优方法。文中重点阐述多假设跟踪算法中数据聚簇、假设生成及假设概率计算、假设约简与剪枝等环节。从工程角度出发,采用K-best最优假设和N-scan回溯剪枝以提高算法效率和实用性。

MHT （Multiple Hypothesis Tracking） is a measurement based on multiple scanning cycles for data connectivity technology. It is theoretically the best method to resolve data association. This article focuses on the multiple hypothesis tracking algorithm for data clustering, hypothesis generation and hypothesis probability calculation, assuming that reduction and pruning and other sectors. From an engineering point of view, the optimal use of K-best hypothesis and the N-scan back pruning to improve the efficiency and practicality.