Generalized Multikernel Maximum Correntropy Kalman Filter for Disturbance Estimation

• 成果类型: Article
• 署名作者: Li, Shilei; Shi, Dawei; Lou, Yunjiang; Zou, Wulin; Shi, Ling
• 署名单位: Hong Kong University of Science & Technology; Beijing Institute of Technology; Harbin Institute of Technology
• 刊物名称: IEEE TRANSACTIONS ON AUTOMATIC CONTROL
• ISSN/ISSBN: 0018-9286
• DOI: 10.1109/TAC.2023.3321368
• 发表日期: 2024
• 页码: 3732-3747
• 关键词: Kernel; Kalman filters; Heavily-tailed distribution; Robustness; CONVERGENCE; Disturbance observers; bandwidth; Generalized loss (GL); multikernel correntropy; robotic manipulator; disturbance observer (DOB)

摘要

Disturbance observers have been attracting continuing research efforts and are widely used in many applications. Among them, the Kalman filter-based disturbance observer is an attractive one since it estimates both the state and the disturbance simultaneously, and is optimal for a linear system with Gaussian noises. Unfortunately, the noise in the disturbance channel typically exhibits a heavy-tailed distribution because the nominal disturbance dynamics usually do not align with the practical ones. To handle this issue, we propose a generalized multikernel maximum correntropy Kalman filter for disturbance estimation, which is less conservative by adopting different kernel bandwidths for different channels and exhibits excellent performance both with and without external disturbance. The convergence of the fixed point iteration and the complexity of the proposed algorithm are given. Simulations on a robotic manipulator reveal that the proposed algorithm is very efficient in disturbance estimation with moderate algorithm complexity.