No Detail Found!

Reduction of computational time for a robust Kalman filter through Leroux gueguen algorithm.

Author(s)
N. Khan Electrical Engineering Department, UET, Peshawar, Pakistan
M. I. Khattak Electrical Engineering Department, UET, Peshawar, Pakistan
T. Bhatti Electrical Engineering Department, UET, Peshawar, Pakistan
Asadulah Electrical Engineering Department, UET, Peshawar, Pakistan
S. R. Shah Electrical Engineering Department, UET, Peshawar, Pakistan
Abstract
Data loss is a frequent dilemma in many processes including state estimation. These lost data samples are normally reconstructed by employing linear prediction theory. Three various linear prediction schemes that includes Normal Equation (NE), Levinson-Durbin Algorithm (LDA) and Leroux Gueguen Algorithm (LGA) may be employed to reconstruct the data loss. The NE method suffers from high computational complexity. On the other hand, LDA is computationally less expensive but it has large dynamic range problem in Linear Prediction Coefficients (LPCs). The LGA overcomes the drawbacks associated with NE and LDA schemes. The major contribution of this paper is the reduction of computational time raised by NE method by employing a modified LGA technique. The upper limit of linear prediction filter order is decided by a minimum mean square error based algorithm. The simulation results are shown by employing this modified LGA on a standard Mass-Spring-Damper system.
Publication Details
Page(s) 91-97
DOI DOI not available
Published Journal: Technical Journal, Volume: 20, Issue: 1, Year: 2015

Request the full text from PASTIC

0

Citations

0

Downloads

116

Views