Intelligent Detection of Guided Scrambling Coded Sequences

Abstract

This document presents a case study in the as-yet unexplored avenue of exploiting running digital sum (RDS) statistics to achieve reduced error rates in the detection of balanced guided scrambling (GS) coded sequences. As demonstrated in this research, balanced GS codes are a good candidate for this approach as the properties of GS coded sequences that produce desirable spectral results translate well to desirable RDS statistics for detection. Through software simulation of GS coded sequences, it is demonstrated that the RDS can be accurately approximated as a cyclostationary Gaussian Markov process. Using an intelligent detection technique developed in this work that takes these RDS statistics into account, error rates over additive white Gaussian noise (AWGN) channels are shown to be significantly reduced. While this research focused on GS coded sequences selected using the minimum square weight (MSW) criteria, the document proposes the development of lower rate codes with high-order spectral-shaping properties that might lend themselves well to this intelligent detection technique achieving even greater reduction in error rates.