Design of an Interleaver for Turbo Codes using Genetic Algorithms
Shoba Paulson, Wim Hordijk, Princy Gift, Subarani, and Shanmugam
International Journal of Artificial Intelligence and Machine Learning 6(2), 1-5, 2006.
Turbo codes are a powerful channel coding technique employed in communication systems. Turbo codes emerged in 1993 and have since become a popular area of communications research. The excellent performance of turbo codes depends strongly on their interleaver pattern. However, finding the best interleaver pattern requires an exhaustive search, and is thus a difficult task. We use a genetic algorithm to search for good interleaver patterns of length N in the space of all possible N! patterns. Compared with random interleavers, the interleaver pattern found by the genetic algorithm is able to achieve a gain of 0.1 db for an interleaver size of N=50.