HARDWARE IMPLEMENTATION OF TURBO CHANNEL CODEC FOR REAL TIME

Abstract

Error control coding is branch of information theory which aims at reducing errors in information from a source to a destination. The concept of error control coding was first introduced in late 1940s by Shannon and Hamming in their hallmark papers [1], [2]. Shannon’s theoretical work defines the upper limit on the achievable data rate of a channel while providing some insights into the general methodology of error control [1]. Hamming, on the other hand, found the first practical code for error mitigation in communication channel and data storage [2]. He actually introduced redundancy into the transmission and did single bit correction in a received bit sequence. Adding redundancy, however, results in decrease of data rate for same system bandwidth or increases bandwidth requirement for same data rate. Decrease in data rate is however compensated by increase in operational range, reduction in error rates and the transmitted power. Soon after Hamming’s work, a plethora of work appeared in the research literature aiming at reducing the error rate for communication or data storage systems. Hamming codes were although maximum distance separable codes, the codes with highest minimum distance for given amount of redundancy, but these suffered from the requirement of certain block lengths. Techniques were then introduced to overcome that requirement by increasing or decreasing the message block length while keeping the minimum distance of the code as large as possible. Elias, in 1955, introduced convolution codes which not only did not suffer from the strict requirement of message block length but also had very simple encoding circuitry [3]. In late 1960s Viterbi introduced decoding algorithm for convolutional codes which changed the focus of research towards convolutional codes [4]. For a very long time convolution codes dominated the wireless communication and data storage systems. It was not until early 1990s, when Turbo codes were introduced, that operating points of codes were near the Shannon limit [5]. Soon after 2 evolution of Turbo codes, techniques to reduce the decoding complexity were introduced resulting in nearly same performance as that of optimal decoder. This thesis aims at describing different decoding algorithms for a configurable Turbo code. These codes are suitable for the systems which can adapt their modulation schemes and code rates depending on the channel condition. Software defined radio is such a system which has the capability to change the data rate by varying code rate and modulation scheme depending on link quality.