Abstract | ||
---|---|---|
For the estimation of real-valued Gaussian signals, the unscented Kalman filter (UKF) can provide a state estimate with second-order accuracy. However, when a general complex-valued system is considered, a direct extension of UKF from the real domain to the complex domain is inadequate, since the complementary covariance information associated with general improper complex-valued signals has been systematically ignored. To this end, in this work, we propose a general complex-valued unscented Kalman filter (GCUKF) algorithm which can be applied for both proper and improper signals. This is achieved by first proposing a novel sigma points selection scheme for the general complex-valued case, followed by a modified state update method to fully utilize both the innovation and its conjugate. A rigorous MSE analysis illustrates the superiority of the proposed state update method, and simulations support the analysis. |
Year | DOI | Venue |
---|---|---|
2022 | 10.1109/LSP.2022.3207414 | IEEE SIGNAL PROCESSING LETTERS |
Keywords | DocType | Volume |
Estimation, Signal processing algorithms, Kalman filters, Technological innovation, Approximation algorithms, State-space methods, Mathematical models, State-space model, general complex-valued unscented Kalman filter (GCUKF), noncircular | Journal | 29 |
ISSN | Citations | PageRank |
1070-9908 | 0 | 0.34 |
References | Authors | |
0 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Xing Zhang | 1 | 0 | 0.34 |
Yili Xia | 2 | 247 | 25.50 |
Chunguo Li | 3 | 48 | 10.72 |
Luxi Yang | 4 | 164 | 22.41 |