Adaptive error correction schemes in channels with transmitter intelligenc

Adaptive error correction schemes are considered for time-varying channels with side information at the transmitter. The transmitter uses the current channel state (side information) to decide the best encoding policy in order to achieve a given decoded error rate specification while providing the greatest throughput possible. These schemes provide a much higher average throughput compared with a conventional scheme that cannot adapt. In most adaptive schemes, the channel state information is not available and can only be estimated by analysing the previously sent messages. By doing this, the channel state estimate is not as accurate as our side information. However, these schemes possess a feedback link that ensures that the encoding-decoding parameters are shared offering high reliability with low complexity. On the other hand, the scenario considered in this thesis is that no feedback link is available, and different techniques for sharing coding parameters with the receiver are presented. These techniques are classified according to whether or not the side information is embedded in the transmission. When side information is embedded the adaptive decoder is quite simple and reliable, but gains in average throughput are modest. On the other hand, when side information is not embedded throughputs can be much greater than those provided by conventional non-adaptive schemes but the decoder strategy becomes complex. Adaptive schemes based on both conventional block codes and punctured convolutional codes are presented and compared in throughput and complexity. As a practical application of the schemes introduced in this thesis, a novel broadcast multiple access communication system is described. Simulation results for this system show that the achievable throughput can be much enhance particularly at low traffic levels when an adaptive strategy is employed.