A Signal Processing Method To Assess Dynamic System Similarity

Market Overview:

This technology features a robust technique to detect dynamic similarity of two systems based on the statistical properties of their signals. The invention assesses the entire autocorrelation structure of a test and a reference signal series. This is achieved in three steps 1) the test and reference signals are subjected to similar pre-processing to guarantee statistical stationarity; 2) the multivariate periodograms or autocovariance functions are calculated for each series; 3) time- and frequency-domain signal discrimination test statistics are computed and assessed. Equality of the test and reference signals is rejected when the multivariate periodograms are too dissimilar and/or sample autocovariance function of the two signals differ greatly.


· System fault diagnosis and similarity assessment of two-time series arising in engineering and non-engineering applications as specified below:

· Engineering: Structural dynamic similarity

· Geophysics: Seismic pattern similarity

· Environmental: Eco-systems dynamics

· Biological: Activity similarity

· Econometrics: Economic dynamics

· Astrophysics: Remote dynamic similarity

· Medical: Diagnostics

· Military: Remote recognition and identification


· Idea is robust and adaptable to many dynamic systems

· System assessment without the need for mathematical (analytical) models

· A novel methodology in signal processing for multivariate (multi-channel) signals

· Outperforms the conventional multivariate signal processing techniques (e.g. Principle Component Analysis)

· Invention presents the first concept in “dynamic recognition” based on observed signals

· Concept includes extension to the time domain to enhance performance


Inventors:                        John Wagner, Hany Bassily, Robert Lund

Protection Status:           Patent issued; # 8,378,816

Licensing Status:            Available for licensing

CURF Ref No:                07-047


Click here for a full color copy with figures

Patent Information:
For Information, Contact:
Chris Gesswein
Director of Licensing
Clemson University Research Foundation
John Wagner
Hany Bassily
Robert Lund
Electrical Engineering
© 2018. All Rights Reserved. Powered by Inteum