A Generalized CHNN Method for Track-to-Track Association

摘要：

A very important aspect of multisensor data fusion is track-to-track association and track fusion in distributed multisensor-multitarget environments. There is a assumption for the proposed approach based on Hopfield neural network that every sensor detect the same targets, but in practice, it is not always realizable. This paper propose a generalized approach based on continuous state Hopfield neural network (CHNN) to solve this problem. Furthermore, the algorithm is generalized to system of three sensors. Also, the Mahalanobis distance is redefined in this paper to accelerate the convergence of the Hopfield networks. Computer simulation results indicate that this approach successfully solves the track-to-track association problem, and it can be generalized in distributed mutisensor-multitarget environment.

关键词： continuous state Hopfield neural network (CHNN) track-to-track association multisensor data fusion.

作者: Baolin He Zheng Mao Yuanyuan Liu Liang Wu

作者单位: School of Electronic Information and Control Engineering,Beijing University of technology,Beijing,100124

会议类型: 国际会议

会议名称: 2009 9th International Conference on Electronic Measurement & Instruments(第九届电子测量与仪器国际会议 ICEMI2009)

会议地点: 北京

会议语种:英文

页码: 4145-4149

在线出版日期: 2009-08-16（万方平台首次上网日期，不代表论文的发表时间）