Implementación de códigos correctores de errores para un módem para la comunicación sobre la red eléctrica

Abstract

The development of turbo codes over the past two decades, has resulted in signi cant performance improvement of PLC communication systems, considerable reduction of the device size, development of more complex chips with channel coding support and implementation of more robust modulation techniques. Today, some of the major research problems are related to the implementation of e cient architectures for decoding algorithms, development of techniques for e cient power consumption management, and latency reduction in the hardware used. The application of turbo codes in communication systems has been characterized for use in di cult conditions for a conventional code. Some of these conditions include the presence of high noise levels in the channel, which require devices to minimize systems power consumption in which the transmission of information has a high cost. In this thesis we present the implementation of the Max-Log-MAP algorithm in TMS320C6416T DSP. We report the results obtained with Gaussian and impulsive noise contaminating the data sent to the receiving device (DSP) as well as the behavior of the decoder for di erent SNR (signal/noise relation). Speci cally the proposal includes the design implementation of the encoder, interleaver, \puncturing" mechanism and decoding data. The implementation presented in this thesis shows the bene ts of a high-performance code in a real application. Based on tests performed on a channel with Gaussian and impulsive noise was observed that increasing the length of the interleaver, the performance increase BER curves. It was further observed that with a turbo encoder type RSC (5,7) the results are very close to the RSC con guration (17,15), also RSC (31,27) encoder generated the best results. On this basis it is possible to reduce hardware resources required to implement a turbo code for the application in the PLC channel