Khmaies Ouahada

Viterbi Decoding of Ternary Line Codes Khmaies Ouahada

The Viterbi Algorithm was originally developed in the 1960's to decode linear error correcting convolutional codes. Its advantages include ease of implementation and using soft decisions to obtain an additional and significant coding gain. Viterbi decoding has found widespread application from space and satellite communications, down to dial up telephone modems and elsewhere, provided that a finite state machine description can be set up. Ternary line codes are used in baseband digital communication systems, notably metallic cable systems. What does the finite state machine description of these widely used nonlinear line codes look like? Can Viterbi decoding be applied to ternary line codes? What gains can be expected with soft decision Viterbi decoding? What advantage can be obtained using more bits for finer soft decision quantization? How do these gains compare to the well known 2 dB gain obtained through 3 bit quantization with linear codes? It is shown that the results obtained differ from the standard 3 bit quantization and 2 dB gain scenario. This study may lead to further new applications of Viterbi decoding with soft decisions.