Advanced decoding of constrained sequence coded signals

点击翻译按钮获取中文摘要

Introduction: Coding is crucial in digital transmission and recording systems in order to ensure accurate communication as throughput, transmission speed, and storage densities increase. Source coding removes redundancy from the data, error control coding enables correction of errors, and constrained sequence (CS) coding helps prevent errors from occurring by avoiding the use of sequences that are prone to distortion on a particular channel. All three forms of coding are ubiquitous in today's high-performance digital communication systems. Objectives: By constraining the form of sequences that can be communicated over the channel, CS codes such as balanced codes, runlength limited codes, and codes that preclude the use of sequences most susceptible to intersymbol interference introduce properties into encoded sequences that improve the performance of the overall system. However, few of the CS decoders developed to date take advantage of the code properties to also intelligently assess the accuracy of the received sequence, instead deferring all error control measures to the error control code. Our initial work investigating advanced decoding techniques for CS codes has shown that it is possible to exploit known characteristics of the CS code in order to increase the accuracy of decoding. The research proposed in this application will extend our initial work to integrate improved sequence detection techniques into CS decoding and therefore provide more accurate data to the error control decoder, resulting in overall increased system performance. Novelty: Our initial investigations have explored two very different approaches for two very different applications:  trellis-based decoding of balanced block codes based on their running digital sum characteristics, and deep learning-based decoding of block codes for visible light communications. In this research we will continue to examine the redundancy inherent in constrained sequences and develop novel approaches that will allow us to exploit this information to increase the accuracy of decoding. This research will also involve the investigation of new CS codes with properties that allow for increasing accuracy of sequence detection, and the use of iterative detection techniques in conjunction with error control decoding. Significance: As transmission speeds and storage densities skyrocket, Canada continues to play a significant role in the development of new communications technologies. This research will enable Canada to remain at the forefront of these technologies so that our economy will continue to benefit from the participation of many researchers and companies involved in this industry. Key outputs of this research will also be the creation of intellectual property to be exploited here in Canada and the education of highly qualified personnel with the skills necessary to step into this high tech industry.