An Improved Particle Filter Based on UKF and Weight Optimization

Aiming at the problem of limited efficiency and accuracy of state estimation in the case of non-linear and non-Gaussian systems, this paper proposes an improved particle filtering algorithm based on edge unscented Kalman filtering and weight optimization for the existing efficiency problems of UPF. Compared with traditional particle filtering, the improved filtering algorithm generates a suggested distribution function in order to avoid excessive variance of particle weights and combines the latest observation information to calculate a more efficient edgeless trace Kalman filter; during the resampling process The weight-optimized resampling method is introduced to solve the problem of particle depletion and improve particle diversity. It can be verified through theoretical derivation and simulation analysis that the improved algorithm effectively improves the calculation efficiency and has better estimation accuracy.