Research On Parameter Estimation Techniques Of Turbo Encoder

The channel coding parameter analyzing is very essential in applications like signal interception and intelligent communication. It uses received coded sequences to analyze the coding parameters, and finally recovers the information sequences by offering parameters for decoding.The Turbo codes realize random coding by introducing interleaver and deinterleaver, and thus achieve near Shannon limit performance. The traditional analyzing methods for Turbo codes usually use hard-decision sequences. These methods have weak fault-tolerance because the reliability information of soft-decision sequences is discarded and their computations are limited in the finite field. However, the Turbo codes are mainly applied in low signal-to-noise ratio (SNR) environment, thus this thesis studies the recognition of Turbo-code parameters by using the soft-decision sequences in low SNR conditions.Focusing on the random interleaver of Turbo codes, an iterative estimation algorithm based on correlation decision between the soft outputs of the Turbo decoder is presented. Firstly, an interleaving matrix is used to describe the permutation relations, and then the estimation of interleaving matrix is converted to the estimation of cross-correlation matrix between the soft outputs of decoder. Considering the estimation reliability, the interleaving matrix estimation is combined with the Turbo decoding, forming an iterative estimation. This algorithm overcomes the shortcomings of existing methods which often aimed at certain interleaving structures only, and suffer from weak fault-tolerance since they use hard-decisions. Simulation results show that the proposed algorithm can recognize any random interleaver in low SNR conditions and has strong fault-tolerance.An estimation algorithm based on the average conformity of parity-check equation is also proposed for the random interleaver of Turbo codes. Firstly, the concept and calculation method of the average conformity of parity-check equation are given, and then the parity-check equations between the information bits and parity-check bits under the unpunctured and punctured conditions are deduced. Then, the permutation positions of the interleaver are estimated step by step, using the truth that the correct permutation position could maximize the average conformity of parity-check equation. Simulation results show that the proposed algorithm can quickly recognize the random interleaver under two different code-rates, and it still performs well in low SNR conditions.As to the double binary convolutional Turbo codes (DB-CTC),two identification algorithms for the code lengths and puncturing patterns of the DB-CTC in DVB_RCS protocol are proposed. One is based on the Maximum a Posteriori (MAP) iterative decoding reliability. Firstly, the decoding reliability is defined for the MAP decoding procedure. The DB-CTC parameters are thus identified based on this concept, using the truth that the correctly separated sequences would maximize the decoding reliability. Then, an algorithm of adaptively controlling the iteration times is proposed for the previous algorithm to improve its efficiency. The other one is based on the average conformity of parity-check equations. Firstly, the parity-check equations between the information sequences and the check sequences under each puncturing pattern are derived. Then, the DB-CTC parameters were identified, using the truth that the correctly separated sequences could maximize the average conformity of parity-check equation. Simulation results show that both proposed algorithms can effectively identify the true code length and puncturing pattern under low SNR for different code rates.