Signal Reconstruction After Noisy Nonlinear Transformations
Abstract
A deterministic signal in zero means Gaussian noise N is observed through a zero memory nonlinearity f(x). The reconstruction of the signal is considered when the nonlinearity, the noise covariance and the first or second order moments of the output process f(s+n) are known. Arbitrary signals can be reconstructed for monotonic and certain odd, not necessarily monotonic, nonlinearities; included here are hard limiters, quantizers and infinite interval windows. Arbitrary signals can be reconstructed, up to a global sign, for two distinct classes of even nonlinearities; included here are 2v-th law devices and symmetric interval windows. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1978
- Accession Number
- ADA060426
Entities
People
- Elias Masry
- Stamatis Cambanis
Organizations
- University of North Carolina at Chapel Hill