State estimation of nonlinear stochastic systems using a novel meta-heuristic particle filter

This paper proposes a new version of the particle filtering (PF) algorithm based on the invasive weed optimization (IWO) method. The sub-optimality of the sampling step in the PF algorithm is prone to estimation errors. In order to avert such approximation errors, this paper suggests applying the IWO algorithm by translating the sampling step into a nonlinear optimization problem. By introducing an appropriate fitness function, the optimization problem is properly treated. The validity of the proposed method is evaluated against three distinct examples: the stochastic volatility estimation problem in finance, the severely nonlinear waste water sludge treatment plant, and the benchmark target tracking on re-entry problem. By simulation analysis and evaluation, it is verified that applying the suggested IWO enhanced PF algorithm (PFIWO) would contribute to significant estimation performance improvements.