Gaussian mixture distributed particle filter algorithm for state propagation using ESPT

This article combines the Gaussian mixture model with the extended propagation single particle filtering technique (ESPT) and proposes a Gaussian mixture distributed particle filtering algorithm that uses ESPT for state propagation. The algorithm is implemented through the average consensus algorithm in both distributed and iterative modes to improve computational efficiency. In this algorithm, the posterior probability is approximated as a Gaussian mixture model, and the fusion of the Gaussian mixture is achieved through importance sampling. Meanwhile, the use of ESPT effectively reduces computational complexity. Numerical examples further demonstrate that compared to other distributed particle filtering algorithms, this method also has significant performance advantages