Research On Encoding And Decoding Alogrithm Of Rateless Codes Over AWGN Channel

With the rapidly development of wireless communication, the mobile communication services demand a higher reliability of information transmission. As one of the important means to ensure the reliability of information transmission, the channel encoding technology is facing with more formidable challenges in complex wireless communication environment and diversified services data transmitting. Especially when mobile communication entered into 3G and developed to 4G, the mobile multi-media services also entered into a stage of rapid development. But the traditional automatic feedback mechanism has been difficult to meet the high efficiency and reliable transmission of broadcast and multicast services of mobile multi-media services in complex wireless communication environment. Based on this, rateless fountain code is proposed for multicast services on erasure channel and widely used in 3G and 4G mobile communication multi-media services. Without requiring channel state information, the transmitter of fountain codes can generate an unlimited number of coding data packages on-the-fly until receives the feedback of decoding successfully from all of receivers of broadcast or multicast, thus to avoid the congestion caused by processing retransmission request from each user. This work researched on the existing channel encoding technology and the encoding and decoding algorithm of fountain code, mainly researched the encoding and decoding optimized algorithm of fountain code under AWGN channel. Based on keeping the original complexity of encoding and decoding, designed encoding and decoding optimized algorithm to fit AWGN channel, used external information transfer(EXIT) figure to optimize degree distribution under specific decoding redundancy cost, explored new technology of fountain code under AWGN channel. The main research work is divided into the following three aspects:Firstly, utilized EXIT to analyze and deduce the relationship between output node degree distribution and input node degree distribution in different decoding redundancies, to obtain the constraint conditions of output node and input node degree distribution during ensuring the rate of decoding successfully, then used linear programming method to optimize the output node degree distribution of fountain code in different decoding redundancy, to obtain the output node degree distribution with the lowest bit error rate under the corresponding decoding redundancy. Thus, in the process of real-time transmission, changing the decoding redundancy and using the corresponding output node degree distribution to reduce the bit error rate and guarantee the reliability of information transmission.Secondly, based on the BER lower bound of fountain code over AWGN channel this work deduced the key factor which influences the decoding performance of fountain code, which is the input node degree distribution. This is because input nodes which are selected few times in encoding, easily due to an encoded packet transmission failure and make the receiver unable recover the entire codes. Base on this, the work proposed a new encoding algorithm. Through selecting input nodes with little degree firstly in each of the encoding process to make the degree of all input nodes equal approximately after encoding. Then union with the optimized output node degree distribution which through EXIT analysis in chapter 2, to encode the original information. The simulation result shows optimize algorithm can reduce the decoding failure probability of fountain code over AWGN channel.Eventually, facing the trend of wireless multi-media services, this work researched the weight fountain code which is proposed recently designed for wireless channels, included the output node degree distribution of decoding performance, optimized the weight probability distribution of weight fountain code to make the encode word values of weight fountain code obey Gauss distribution after encoding. Meanwhile, applied the optimized output node degree distribution from chapter2 and encoding algorithm from chapter3 to weight fountain code and obtained the performance optimization of coding and decoding.