Statistical convergence analysis of Rosenblatt's perceptron algorithm as a DS-spread spectrum detector

摘要

A stochastic analysis is presented for the learning behavior of a single-layer perceptron when used as a direct sequence (DS) spread spectrum detector. The input is a noisy DS-spread spectrum BPSK signal, the training data is a binary sequence, and the perceptron weights learn using Rosenblatt's (1962) algorithm.