Research And Application Of Distributed Coding And Decoding And Variable-length Interleaver For Turbo Code

Turbo codes was proposed in1993, by C.Berrou. Its excellent anti-interference performance makes it subject to attention. Convolutional codes and random interleaver are combined to achieve random coding. Decoding algorithm is soft output iterative decoding algorithm, witch is tried to approximate the maximum likelihood decoding. The conditions of Shannon noisy channel coding theorem are considered fully. Therefore, the excellent performance of just under Shannon limit0.7dB is obtained for Turbo codes.Turbo codes can achieve excellent performance, but it also contain some shortcomings, such as decoding delay, high complexity of computation and large capacity of memory etc. In this paper, the key technology and working principle of Turbo codes are analyzed. The iterative decoding process and the decoding algorithm of Turbo codes are studied deeply, and then, their performances are compared and analyzed by simulation. In addition, the basic interleavers of Turbo codes are designed in this paper. On the basis of S-random interleaver, one new kind of interleaving algorithm is proposed. Finally Turbo codes for wireless sensor networks are studied.The innovations of this paper are as follows: Firstly, Interleaving is one of the key technologies for Turbo codes to achieve outstanding performance. Interleaver is introduced to reduce the correlation between the parity bits. It is helpful to achieve random coding for Turbo codes. Thus, the bit error rate in the iterative decoding process is reduced. Via studying the basic interleaver of Turbo codes, we found that the S-random interleaver has the best performance, but when the interleaving length is longer or parameter S increases, the computational complexity is too high to be realized. Based on the S-random interleaver, this paper proposes a new interleaver, which makes use of several short length S-random interleavers to construct a longer length interleaver. The simulation indicates that the computational complexity is reduced significantly, while the performance is almost good as the S-random interleaver’s. And with the increase of interleaving length, the bit error performance is improved. The proposed interleaver has robustness against pruning of the interleaver length.Secondly, in the wireless sensor networks, a suitable channel coding scheme is indispensable to protect the transmission of information to achieve the desired bit error rate, while ensuring that the sensor node power consumption and computational complexity are low. Based on the traditional Turbo codes, this paper proposes a distributed algorithm for turbo encoding and decoding. The sensor nodes are assumed in neighborhood of a source, thus, their observed data is correlated. The decoding algorithm is distributed low-complexity iterative decoding algorithm and is derived in detail in this paper. By anglicizing the simulation, it is known that the proposed distributed Turbo codes has better performance on the micro-sensor nodes with limited hardware resources.