Post by xuke on Feb 26, 2009 1:01:59 GMT -5
I am doing my Ph.D. research on turbo equalization. As known the turbo equalizer operates like the serial turbo codes, the equalizer is like the inner decoder and the decoder in turbo equalization is like the outer decoder.
But I am still a little confused about the serial turbo codes. Usually the SISO decoder has two inputs: the prior probabilities and the channel values, and one output: the extrinsic information. For the serial turbo codes, the outer decoder does not have the access to the channel values and only operates on the likelihoods produced by the inner decoder. Then two choices for the outer decoder can be given intuitively. (1) All the prior probabilities for the outer decoder are assumed to same (or zeros for logarithm), and using LLRs from the inner decoder as its channel value. (2) All the channel values for the outer decoder are set to zeros, using the LLRs from the inner decoder as its prior probabilities.
I am not sure which method is the exact one the serial turbo codes use. Could you help me with this problem? Thanks a lot.
But I am still a little confused about the serial turbo codes. Usually the SISO decoder has two inputs: the prior probabilities and the channel values, and one output: the extrinsic information. For the serial turbo codes, the outer decoder does not have the access to the channel values and only operates on the likelihoods produced by the inner decoder. Then two choices for the outer decoder can be given intuitively. (1) All the prior probabilities for the outer decoder are assumed to same (or zeros for logarithm), and using LLRs from the inner decoder as its channel value. (2) All the channel values for the outer decoder are set to zeros, using the LLRs from the inner decoder as its prior probabilities.
I am not sure which method is the exact one the serial turbo codes use. Could you help me with this problem? Thanks a lot.