Return to search

Channel characterisation and coding for the FM SST channel

M.Ing. / The transmission of digital data at higher data rates and greater reliabilities is becoming increasingly more important in the society of today. The usage of, especially wireless digital data, is increasing at an alarming rate, and the need arises for more bandwidth to be made available for this purpose. Hence, the need arises to optimise and expand the usage of bandwidth currently under use. Existing wireless channels, such as the commercial FM service, need to be re-evaluated and unused bandwidth needs to be developed and used as close to capacity as possible. In this dissertation, a study was launched to develop the FM SST (Supplementary Signal Transmission) commercial analog service as a simplex digital channel that can commercially be used to transmit a myriad of digital data. With the help of a commercial FM radio channel, digital data was transmitted using a number of different modulation schemes to obtain channel measurements for the FM SST channel. The modulation scheme that performed the best, given certain criteria, was GMSK. This raw information was then used to extract first and higher order statistical information from the channel, in order to parameterise discrete channel models, based on a Markov process. The channel models that were used were the Gilbert, GilbertElliott and Fritchman channel models. These channel models were then tested according to their iirst and higher order statistical data to evaluate their effectiveness in modelling the error process on the FM SST channel. It was found that the channel model best approximating the real channels was the Gilbert channel. Once satisfactory results were obtained, these discrete channel models were then used to evaluate various error correction schemes for their ability to correct burst errors on the FM SST channel. From the various schemes evaluated, a (63,39) BCH with an interleaver of index 4 was found to perform the best. The purpose of which is to transmit digital data at the highest data rate possible and at the same time having a bit error rate less than lxl0-6.

Identiferoai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:uj/uj:7269
Date10 November 2011
CreatorsAlbertyn, Eben
Source SetsSouth African National ETD Portal
Detected LanguageEnglish
TypeThesis

Page generated in 0.0025 seconds