分布式多目标跟踪中数据融合的一种新方法

A New Method for Data Fusion in Multitarget Tracking

数据融合问题是分布式多传感器多目标跟踪中的关键问题。相关文献在航迹和假设的相容性判别以及数据融合过程的分析等问题上存在一些不足。该文从“无矛盾原则”出发 ,给出了准确而完善的航迹和假设相容性判别条件 ,澄清了前人在此问题上的一些模糊概念。对航迹和假设的融合过程做了详细的分析 ,提出了“定向融合航迹”和“自由融合航迹”的概念。按照新的航迹和假设相容性判别条件 ,进行了仿真实验。

Data fusion is very important in multitarget tracking with a distributed sensor network. Fusable conditions of traces and hypotheses presented in related literature are not so precise and the fusion process of traces and hypotheses is not discussed very clearly.Based on the principle of 'no contradiction',precise and complete fusable conditions are obtained. The fusion process of traces and hypotheses is analyzed in detail.Two concepts called 'fixed fusing trace' and 'free fusing trace' are presented. Simulation results show that the new fusable conditions can work properly.