Convergence Analysis and Improvement of Progressive Gaussian Approximation Filters

This paper studies the convergence problem of progressive Gaussian approximation filters (PGAFs). The convergence analysis of the PGAF is presented based on the Lyapunov method. It is proved that the estimation error is bounded when the convergence conditions hold. Based on the convergence analysis, the filtering performance of the PGAF is improved by adjusting the strategy of the progressive measurement update. Finally, a simulation experiment of mobile target tracking is designed to verify the effectiveness of the analysis results, and it demonstrates that the improved PGAF method has higher than the existing methods