MCMC for Sequential Flight Object Attitude Estimation Based on Perfect Coupling Sampling
Aiming at large initial attitude errors of flight object, this paper presents perfect coupling sampling based on coupling from the past (CFTP) algorithm on MCMC (Markov Chain Monte Carlo) to tackle the problem of sequential flight object attitude estimation. Based Bayesian theory, posterior distribution can be approximated by Monte Carlo likelihood function and conjunction prior distribution based via constructing MCMC of flight object monotonous state-space and difference encoding. It is so-called perfect-sampling of MCMC method, which can guarantee that samples are drawn exactly from distribution of flight object attitude estimation. Simulation results show that this method can reduce computing complexity and effectively explore the time of convergence of sequential flight object attitude estimation.
Perfect coupling sampling Flight object attitude estimation Coupling from the past (CFTP) Stationary distribution
Zhang Jingmei Zhai Yongzhi
School of Automation,Northwestern Polytechnical University,Xian 710072,China School of Electronics and Information,Northwestern Polytechnical University,Xian 710072,China
国际会议
深圳
英文
1418-1421
2008-12-10（万方平台首次上网日期，不代表论文的发表时间）