Research On Automatic Recognition Of Channel Codes Based On Soft Information

With the rapid increasement of users in mobile communications, the channel resources become increasingly scarce. To alleviate this problem, adaptive modulation and channel coding is introduced to adaptively select the modulation and channel codes according to channel quality, and transmit the corresponding information through the control channel. However, due to some time-varying factors such as noise, interference, delay, the receiver may obtain the control information non-timely and inaccurately. Meanwhile, frequently transmitting control information makes the overhead of the control channel obvious increasement. Therefore, the automatic recognition of channel codes is proposed, which can recognize the channel code of the transmitter by the soft information of the received sequence when candidate channel codes set is known.In addition, in the field of non-cooperative information capture, the non-cooperative receiver does not accurately know the channel code of the transmitters, but it can conjecture a set of channel codes utilized by the transmitters, via a priori information or empirical knowledge. In such a scenario, by the soft information of the captured sequence, the automatic recognition of channel codes can be applied to determine the channel encoding used in transmitters, which lays a foundation for the subsequent channel decoding.Thus it can be seen, the automatic recognition of channel codes is significant for cooperative mobile communications and non-cooperative information capture. This thesis will focus on the method or algorithm for automatic recognition of channel codes based on soft information obtained by the demodulated signal in the receiver. The main work is as follows:1. the method for automatic recognition of convolution codes, linear block codes and LDPC codesAn improved algorithm for automatic recognition of convolution codes, linear block codes and LDPC codes is proposed. The improved algorithm firstly calculates the posterior probability log-likelihood ratio by the soft information of the received sequence, and then uses the weighted average of it as recognizing characteristic, finally find the maximum of recognizing characteristics to make a decision. Theoretical analysis shows that, compared with the existing algorithm, the improved algorithm can effectively reduce the complexity and is easy to implement. Simulation results demonstrate that the improved algorithm can achieve higher detection probability in a short detection length and meet the requirements of practical systems.2. the method for automatic recognition of Turbo codesAn algorithm for automatic recognition of Turbo codes is proposed. Based on the structure of Turbo codes, the proposed algorithm firstly demultiplexes and depunctures the soft information of the received sequence to attain the soft information of two sequences corresponding to the output of component encoders. Then the soft information of punctured bits is patched. Finally, the method for automatic recognition of convolution codes is used to recognize the component encoders, puncture structure and interleaver. Consequently, the Turbo codes used in transmitters is recognized. By simulations, the effectiveness of the proposed algorithm is verified.