HF communication has been used for more than a century and to this day still fulfils an important function in communications networks. In order to interface with modern communications protocols, solutions have to be designed to facilitate data communication over HF (High Frequency). STANAG 5066 is one such solution which provides an application independent ARQ (Automatic Repeat Request) bearer service for client applications. A need exists within the STANAG 5066 specification for a DRC algorithm. The objective of such an algorithm is to select the optimum data rate and interleaver size, based upon current HF channel conditions, to maximise the data throughput over the HF link. In this dissertation previous implementations of DRC algorithms were studied and evaluated. In literature it was found that algorithm implementations used the FER and no channel information to make a data rate choice. This resulted in algorithms that tended to oscillate between data rate choices, and was very slow to react to changes in the HF channel. A new DRC algorithm was designed and simulated that uses the SNR (Signal-to-Noise Ratio) and the BER estimate to make a data rate choice. The DRC algorithm was implemented in a commercial STANAG 5066 system and tested using HF data modems and a simulated HF channel. The results of the implementation and testing show that the designed DRC algorithm gives a better performance, is quicker to adapt and is more robust than previous DRC algorithms. This is also the first DRC algorithm that has been designed to use channel information, such as the SNR and BER, to make a data rate choice.
Dissertation (MEng)--University of Pretoria, 2007.