Conference Record of the Thirty-First Asilomar Conference on Signals, Systems and Computers (Cat. No.97CB36163) (1997)
Pacific Grove, CA, USA
Nov. 2, 1997 to Nov. 5, 1997
G.A. Hanson , Weapons Div., Naval Air Warfare Center, China Lake, CA, USA
We consider the problem of detecting complex signals with uncertain parameters including amplitude and phase. For one class of problems the optimal solution is a three-layer neural network with an infinite number of intermediate nodes. We investigate several finite size structures which approximate the output of the optimal detector and deliver near optimal detection performance with reduced complexity. Training these structures is shown to have an interpretation in terms of minimizing cross entropy.
signal detection, optimisation, Gaussian distribution, neural nets, minimum entropy methods, approximation theory
G. Hanson and R. Iltis, "Near optimal detection of complex signals with unknown parameters," Conference Record of the Thirty-First Asilomar Conference on Signals, Systems and Computers (Cat. No.97CB36163)(ACSSC), Pacific Grove, CA, USA, , pp. 530-534.