Key words: Joint Source Channel Trellis Coding, BCJR, MAP, Viterbi, Pairewise Error Probability
Abstract: A Joint Source-Channel Trellis Coding technique consisting in the joint optimizaton of a trellis source coder and a convolutional code is presented. The pairwise error probability needed for the codebook design and the encoding processes is estimated theoretically with the transfer function of the convolutional code. This way, no extensive monte-carlo simulations are required. In addition, at the receiver end the BCJR algorithm is modified to act as a "soft source decoder". By using the channel estimation of the BCJR algorithm, the distortion introduced by channel errors can be further reduced.Simulation results for different benchmarking sources are presented.
Full Paper: Click Here
Authors: Emmanuel Boutillon, Luis Gonzalez Perez
Reference: Proceedings of the 2nd International Symposium on Turbo Codes & Related Topics, pp 415-18, Brest, France, Sept. 2000.