Deep Learning-based Blind Estimation for the Number of Users in Multi-User DSSS Systems

Abstract

In non-cooperative contexts, a receiver must estimate the transmitter's communication parameters to recover information. Particularly in multi-user direct sequence spread spectrum systems, it is crucial to estimate the number of users and determine the length and pattern of each spreading sequence. Various studies estimating the spreading sequence have assumed that the number of users is known beforehand. And, the methods employed by these studies in an attempt to justify this assumption have inherent issues, such as peak ambiguity and threshold setting. To address these issues, this paper proposes a deep learning-based blind estimation algorithm for determining the number of users using the autocorrelation fluctuation extracted from the input signal. Through computer simulations, we show that the proposed method achieves high estimation accuracy even at low signal-to-noise ratio.