Conditionally bilinear filter with tracking application |
| |
Authors: | WJ Kolodziej RR Mohler |
| |
Institution: | Department of Electrical and Computer Engineering, Oregon State University, Corvallis, OR 97331, USA U.S.A. |
| |
Abstract: | A recursive nonlinear filter and tracking methodology is developed here for a class of partially observable processes. The method is based on an approximation of a nonlinear system by a system which is linear in the unobservable states and has the unobservables conditionally Gaussian with respect to the observations initially. Model smoothness, such as required with most approximating filters is not required here, and applications to simulated tracking problems show the filter to be considerably more accurate than the modified second-order filter which in general sense includes the extended Kalman filter. |
| |
Keywords: | |
本文献已被 ScienceDirect 等数据库收录! |
|