On multi-Bernoulli approximations to the Bayes multi-target filter

摘要：

Mahler recently proposed the Multi-target Multi-Bernoulli (MeMBer) recursion as a tractable approximation to the Bayes multi-target recursion, and outlined a Gaussian mixture solution under linear Gaussian assumptions. These proposals are speculative in the sense that, to date, no implementations have been reported.In this paper, it is shown analytically that the MeMBer recursion has a significant bias in cardinality that results in a high number of false tracks. A novel approximation that alleviates the bias problem is proposed. In addition, a sequential Monte Carlo implementation (for generic models) and a Gaussian mixture implementation (for linear Gaussian models)are given. Comparisons with Mahlers original MeMBer filter via simulations show significant reduction of false tracks.

关键词： small Tracking estimation random sets point processes finite set statistics multi-Bernoulli

作者: B.-T.Vo B.-N.Vo A.Cantoni

作者单位: University of Western Australia.Crawley, WA, Australia EEE Department University of Melbourne Parkville, VIC, Australia University of Western Australia Crawley, WA, Australia

会议类型: 国际会议

会议名称: The International Colloquium on Onformation Fusion 2007(2007年国际信息融合研讨会)

会议地点: 西安

会议语种:英文

页码: 26-36

在线出版日期: 2007-08-22（万方平台首次上网日期，不代表论文的发表时间）