Expectation-maximization algorithm for direct position determination

详细信息    查看全文

• 作者：Elad Tzoreff ^{eladtzor@post.tau.ac.il} ; Anthony J. Weiss ; ^{ajw@eng.tau.ac.il}
• 关键词：Time of Arrival (TOA) ; Expectation Maximization (EM) ; Laplace Method ; Direct Position Determination (DPD)
• 刊名：Signal Processing
• 出版年：2017
• 出版时间：April 2017
• 年：2017
• 卷：133
• 期：Complete
• 页码：32-39
• 全文大小：1109 K
• 卷排序：133

文摘

A framework to solve TOA/TDOA based localization in the presence of nuisance parameters, using the EM algorithm. Application of the framework to the DPD method which is usually solved using variants of exhaustive search, due to the non-convex structure of the objective function. Demonstration by simulations that the proposed EM approach achieves the global maximum at various scenarios, and SNR levels with high probability. Application of the EM algorithm to the traditional two step method with a closed from solution at each iteration without using derivatives (or Jacobians). Demonstration that the algorithm is more stable than the well-known Gauss-Newton algorithm, and generally converges quickly.