Some Design Considerations on the Adaptive Quantizer Detector in Unknown Noise.

摘要

The contents of this report provide some general guidelines and insights into the design of an adaptive signal detection system based on quantized data. In particular, we focus on the adaptation of the parameters in the quantizer in a way such that the detector's probability of making errors in the presence of noise of unknown statistics will be minimized with respect to those parameters. The unknown noise is assumed throughout this report to be an independent, additive noise with a symmetric (about zero) probability density function. Within this assumption an adaptive scheme is developed, and its performance and convergence are verified via simulation. (Author)