Symbol-Level Synchronisation Channel Modelling With Real-World Application: From Davey-Mackay, Fritchman to Markov

TL;DR

This paper introduces a novel error category-based channel model incorporating insertion, deletion, and substitution errors, validated with real visible light communication data, highlighting the channel's inherent memory and improving realism over traditional models.

Contribution

It extends existing models by incorporating synchronization errors with memory, using real data, and introduces a new parameter estimation method based on Levenshtein distance.

Findings

The new model deviates from the memoryless Davey-Mackay model.

It captures the inherent memory in synchronization channels.

The method is applicable to other fields like DNA barcoding.

Abstract

Errors in realistic channels contain not only substitution errors, but synchronisation errors as well. Moreover, these errors are rarely statistically independent in nature. By extending on the idea of the Fritchman channel model, a novel error category-based methodology in determining channel characteristics is described for memory channels which contain insertion, deletion, and substitution errors. The practicality of such a methodology is reinforced by making use of real communication data from a visible light communication system. Simulation results show that the error-free and error runs using this new method of defining the channel clearly deviates from the Davey-Mackay synchronisation model which is memoryless in nature. This further emphasises the inherent memory in these synchronisation channels which we are now able to characterise. Additionally, a new method to determine the parameters of a synchronisation memory channel using the Levenshtein distance metric is detailed. This method of channel modelling allows for more realistic communication models to be simulated and can easily extend to other areas of research such as DNA barcoding in the medical domain.