Reaseach On Convolutional Code Based On3G

With the continuous progress of communications technologies, some specifications for communications system have become increasingly significant. How to improve the information transmission rate in the interference environment has been a hot research field in recent decades. As an important technology to improve the reliability of information transmission, error-correction code improves the information transmission rate by reducing the number of times of retransmission of the information. Combining with Viterbi decoding algorithm, convolutional codes have the advantages of low decoding delay and excellent error-correction capability, which makes convolutional code widely used in modern communication systems.Firstly, the basic principle of convolutional codes and its distance properties are given in this thesis. After a brief description of the three main technical standards for3G, the application of convolutional codes in WCDMA system has been analyzed. Then, elaborated convolutional codes with Viterbi decoding algorithm, and detailed analysis of the key parameters that affect its performance are conducted.Secondly, this thesis focuses on exploring the structural properties of convolutional codes. Based on the concept of primitive polynomial in modern algebra theory, some kinds of primitive convolutional code are conceived. Subsequently, the performance simulation corresponding to primitive convolutional code and non-primitive convolutional code by using MATLAB is provided. The simulation results show that both the constraint length and code rate have the same impact on the performance of these two kinds of convolutional code. Meanwhile, after obeserving and analysing multiple sets of data it is found that the soft decision Viterbi algorithm has2dB coding gain or so compared with the hard decision Viterbi algorithm when the BER=10-6.Moreover, the comparison between the configured primitive convolutional code and non-primitive convolutional code in the performance graph demonstrates that, with the constaint length K increasing, the performance between these two kinds of convolutional code changes again. The performance of the primitive convolutional code is signigicantly better than the non-primitive convolutional code when constraint length K=3, in the case of K equals7or8, and the performance of primitive convolutional code is still overmatch the non-primitive convolutional code when the signal-to-noise ratio is less than1.5dB; However, with the increasing of the signal-to-noise ratio, the error-correction performance of non-primitive convolutional codes outperforms the primitive convolutional code.Finally, the non-primitive convolutional code emplored in WCDMA system is explored, which has the maximun free distance and theoretically be the best error-correcting code among all the convolutional codes with the same constraint length. However, the constructed primitive convolutional code with the same constraint length, has better error-correction performance when signal-to-noise ratio less than1.5dB. Theoretical analysis of the simulation results shows when the signal-to-noise ratio is low the maximum free distance is no longer the dominant factor affecting the error-correction capability of convolutional code. Correspondingly, the constructed primitive convolutional code in this thesis has a specific structure, containing a small number of low weight codeword sequences, which make its error correction capability in low signal-to-noise ratio better than that of the maximum free distance.