Optimum Predictive Iterative Decoding of Turbo Codes in Coloured Gaussian Noise

K. Vasudevan (India)


Turbo codes, prediction filter, coloured noise.


This paper addresses the problem of iterative decoding in the presence of coloured noise. Although most of the current literature deals with iterative decoding in additive white Gaussian noise (AWGN), in many practical situa tions, the noise is correlated. This correlation is usually due to the use of a linear equalizer to remove ISI. In this pa per, we first derive the optimum detector for coloured noise and then extend the concept to iterative decoding. Simula tion results show that depending on the noise correlation, performance improvement of the order of 511 dB can be obtained by using the predictive iterative decoder.

Important Links:

Go Back