Particle Filter with Weight Adjustment and Its Application on Heat Exchange Process with Time Delay and Large Sensor Noise
Yuanhong Xu, Mingcong Deng · 2022
Particle filter has been popularized and used for solving estimation problems since the early 1990s. Extensive research has advanced the generic particle filter to improve the performance and applicability in various ways. A variant of the generic particle filter algorithm is shown in this paper. The particle filter variant re-adjusts the calculation method of the particle weight during the update stage, where the normalized exponential weight instead of the original absolute particle weight is used to deal with the problem of unavailability of reliable weights when encountering large realistic noises. Simulation shows the effectiveness of the method.