The non-recursive estimation using a batch filter based on particle filtering

In this paper, the non-recursive estimation algorithm using a batch filter based on particle filtering is developed and utilized for one-dimensional nonlinear example. For comparison study, algorithms of a batch filter based on unscented transformation and generic particle filtering are briefly reviewed and new algorithm of a batch filter based on particle filtering is presented. For verification of presented batch filter´s performance, the numerical simulations and the accuracy assessment are achieved and results are compared with those of a batch filter based on unscented transformation for various nonlinear and non-Gaussian environments. The root mean square value of differences between observational values and computed values after convergence is used for precision check of filtering process, estimated initial state value and the difference between true initial state value and estimated initial state value are used for state accuracy check of nonlinear estimation. Large initial state and various type of measurement noise are used for nonlinear and non-Gaussian environments, respectively. Under large initial state error or large non-Gaussian measurement noise, the developed non-recursive estimation algorithm give more robust and accurate estimation results than those of a batch filter based on unscented transformation easily. Finally, the non-recursive batch filter based on particle filtering is effectively applicable for batch estimation problems under nonlinear and non-Gaussian environments.