Multipath Estimation Algorithm Based on Improved Unscented Kalman Filter
ZHANG Jinheng,CHENG Lan,ZHANG Jing,NI Zihang,YAN Gaowei
DOI: https://doi.org/10.16355/j.tyut.1007-9432.2023.05.016
2023-01-01
Abstract:Purposes In navigation and positioning system, the multipath estimation algorithms based on Kalman filter framework can effectively improve the positioning accuracy. When the initial value of the process noise and observation noise covariance of such algorithms is improperly selected, a large error or even divergence of the estimation results may occur. In addition, because the algorithm is based on the minimum mean squared error criterion, it is susceptible to non-Gaussian noise, especially under heavy-tailed non-Gaussian noise, which has the problem of significant degradation of estimation accuracy. Methods In order to maintain good multipath estimation results under both Gaussian noise and non-Gaussian noise and improve positioning accuracy, an adaptive maximum correntropy unscented Kalman Filter (AMCUKF) multipath estimation algorithm is proposed in this paper. The AMCUKF algorithm introduces the maximum correntropy as an optimization criterion in the process of observation update to solve the problem of estimation accuracy degradation under non-Gaussian noise. In the process of noise covariance update, the residual sequence of the observed quantity is used to recursively update the noise covariance to solve the improper initial value selection of the process noise and the observed noise covariance. Findings Simulation experiments are carried out under Gaussian noise and non-Gaussian noise, and by comparing with two estimation algorithms based on Kalman filter framework, it is shown that AMCUKF multipath algorithm can not only maintain better multipath estimation results under Gaussian noise, but also maintain higher multipath estimation accuracy under non-Gaussian noise, effectively suppressing the interference of non-Gaussian noise.