Neural net based variable structure multiple model reducing mode set jump delay

Variable structure multiple model (VSMM) is one of the most powerful algorithms for effectively tracking a single maneuvering target. Although VSMM is developed specifically to improve the interactive multiple model (MM) method focused to reducing computational cost and improving tracking performanc...

Full description

Saved in:
Bibliographic Details
Published inProceedings of the 11th IEEE Signal Processing Workshop on Statistical Signal Processing (Cat. No.01TH8563) pp. 142 - 145
Main Authors Daebum Choi, Byungha Ahn, Hanseok Ko
Format Conference Proceeding
LanguageEnglish
Published IEEE 2001
Subjects
Computational efficiency
Delay effects
Estimation error
Filtering
Filters
Markov processes
Mechatronics
Neural networks
Noise measurement
Target tracking
Online AccessGet full text

Cover

More Information
Summary:Variable structure multiple model (VSMM) is one of the most powerful algorithms for effectively tracking a single maneuvering target. Although VSMM is developed specifically to improve the interactive multiple model (MM) method focused to reducing computational cost and improving tracking performance, it presents an inherent limitation in the form of the presence of mode set jump delay (MJD). MJD as an undesirable phenomenon in VSMM is described and analyzed. In order to eliminate the MJD, a neural network based VSMM that automatically selects the optimal mode set as achieved by supervised training is proposed. Through representative simulations we show the proposed algorithm outperforming over the conventional digraph switching VSMM in terms of tracking error.
ISBN:9780780370111
0780370112
DOI:10.1109/SSP.2001.955242