Error control coding combined with content recognition

In modern communication systems, there is content redundancy inside the transmitted data no matter what source codes are adopted. This paper introduces machine learning method to utilize content redundancy and perform content recognition, in order to provide side information for error control coding (ECC) and improve performing. It shows that content recognition for English text, based on Trie and machine learning, is able to provide side information for ECC decoders. Moreover, content redundancy is independent of parity bits of error control codes. Iterative Turbo decoding is introduced to further enhance error performance. Simulation results show that the error performance of ECC can be significantly improved with the aid of content recognition.