Denoising technique based on cascaded filtering of particle filter and ANFIS

For the practical application of nonlinear, non-Gaussian noise system state estimation, this paper develops the ANFIS- Particle filter cascaded filtering model based on the adaptive neuro-fuzzy inference system(ANFIS) nonlinear approximation function and particle filter's obvious advantages for non-linear state estimation. ANFIS is used to eliminate the bias in the colored noise of the signal, then the filtered signal is processed by the particle filter to realize the optimal state estimation. The simulation results demonstrate that with the cascade filter model the mean and variance are reduced by 65% and 74% respectively, ANFIS-particle filter model has significant noise cancellation effect for strongly nonlinear systems, and the state estimation accuracy has been greatly enhanced, which verifies the effectiveness of the proposed model.