.png){kind=icon}
.png){kind=icon}
Return
Adaptive Estimation Using Interacting Multiple Model With Moving Window
.png){kind=icon}
OA
AI
DOI：10.1109/ACCESS.2024.3422255.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
State estimation is paramount in control, monitoring, and fault management across various domains. Uncertainty in model parameters and changing system dynamics pose significant challenges to accurate state estimation. This paper proposes a novel adaptive estimation strategy called the Moving Window Interacting Multiple Model (MWIMM). Using a moving window improves identifiability and computational efficiency of the multiple model algorithms by focusing on a subset of possible models, rather than considering all models at each stage. MWIMM enables the estimation of gradual changes in the system, making it valuable for fault intensity and Remaining Useful Life (RUL) estimation. The paper provides an overview of adaptive estimation strategies, presents the formulation of MWIMM for fault intensity and RUL estimation, and investigates the parameter estimation problem. Results are compared with those of augmented state Extended Kalman Filter (EKF) estimation, and it is shown that the proposed MWIMM approach offers a promising alternative for effectively handling extensive parameter uncertainty and accommodating gradual changes in system parameters.
Keywords:
Adaptive estimation
moving window IMM
fault diagnosis
RUL estimation
moving window IMM
fault diagnosis
RUL estimation