Research On Interleaver And Iterative Decoding Stop Criterion For Turbo Code

Turbo code is a new channel coding technique. It brings a revolution for the channel coding theory and practice, especially with the incomparable advantage of traditional channel coding technology in anti fading and interference. Making it the INMARSAT-phone M4 system, CDMA multiuser detection and third-generation IMT-2000 mobile communication system has been widely used. But Turbo code also has some problems to be solved. People need further study to solve it. This thesis actively studied interleaver and iterative decoding stop criteria of Turbo code, as follows:Firstly, this thesis analyzes the basic principle of Turbo coding and decoding, and also introduced LOG-MAP and SOVA decoding algorithm. On this basis, the parameters affecting the performance of Turbo codes (i.e., the component codes, interleaver, puncture, a number of iterations, and the decoding algorithm) were simulation by Mathlab software. And the thesis gives reasonable parameters to improve the performance of Turbo codes.Secondly, by analyzing the working principle of typical interleaver, a new semi-regular interleaver is designed for the disadvantage of spiral interleaver. The new interleaver uses random array to rearrange the order of interleaving matrix ranks and read out data with a positive spiral way. And making sequence has mode 2 features. It not only increases the distance between the adjacent bit-interleaved, and solves the problems of protection of the information bits after the unevenness caused by the puncturing matrix. In order to achieve the best semi-regular interleaver working performance, this thesis use genetic algorithms to search the array of the matrix and find the best matrix ranks. The simulation results show that the semi-regular interleaver is better than the performance of bloke interleaving, spiral interleaver and random interleaver, and has the same performance of S-random interleaver. After optimization, the performance of semi-regular interleaver is maintained in the optimal state, the bit error rate at the same SNR is also obtained a great gain.Finally, in order to reduce the decoding delay and the average decoding iterative times, the iterative decoding stop criterion is studied. Improved HDA criterion and improved SDR criterion have been proposed. Improved HDA criterion defines the amount of unreliability based on the external information. When this amount of unreliability does not meet the threshold, it stops decoding. Improved SDR criterion uses the ratio of the number of difference between two successive iterations output symbols to replace the original function, so as to solve the problem of the original criterion cannot stop decoding when the frame is not convergence. The simulation results also show that the average number of iterations to the two improved criteria than the original criterion has been significantly reduced.