Title :
Optimal linear associative memories for noise rejection
Author :
Zou, Li-He ; Lu, Jun
Author_Institution :
Dept. of Electr. Eng., Louisiana Tech. Univ., Ruston, LA, USA
Abstract :
An approach to design of linear associative memories for rejection of colored input noise is proposed in the sense of the least mean square error. A null-space projection model, an optimal unbiased model and an optimal biased model are developed. Computer simulation results demonstrate that the proposed models have a superior noise rejection performance over the Kohonen model. Especially, the optimal unbiased and the optimal biased models show excellent efficiency
Keywords :
content-addressable storage; interference suppression; noise; colored input noise; least mean square error; linear associative memories; noise rejection; null-space projection model; optimal biased model; optimal unbiased model; Associative memory; Colored noise; Computer simulation; Covariance matrix; Ear; Least squares approximation; Mean square error methods; Symmetric matrices; Vectors; White noise;
Conference_Titel :
Circuits and Systems, 1992., Proceedings of the 35th Midwest Symposium on
Conference_Location :
Washington, DC
Print_ISBN :
0-7803-0510-8
DOI :
10.1109/MWSCAS.1992.271084
