Design And Research On The Algorithms Of Convolutional Network Error Correction Encoding And Decoding

As a new technology field in network communication theory,the core idea of Network Coding is that when information transmitted over the network,its intermediate nodes can make encoding processing of the coming information and forward.Compared to the traditional store and forward mechanism,network coding has advantages in throughput,data security,robustness and load equalization and so on.In practice,due to the delay of information and cycles existing in network,the messages from different times convolves together,which naturally causes convolutional network coding.So it is more significance to research convolutional network coding.In convolutional network,due to human factors and a variety of noise,information will occur kinds of errors in network transmission.Therefore,it is necessary to study the error control in convolutional network coding.The problem of error correction in convolution network coding is studied in this master.At present,there are two main directions in research of network error correction coding which one is based on the classical error correction coding theory and another is based on the rank distance theory.The former achieves the purpose of correcting the error by increasing the redundancy in the time domain,while code words of the latter is represented by subspaces,and the error control is realized by introducing the redundancy in the space domain.We do in-depth research both in those two methods and have achieved certain results.The principle content and contributions of this master include the following two aspects.Convolutional network theory is the point of penetration in the first part of this master.We combined the classical convolutional code error correction technique with convolutional network coding and designed an encoding and decoding method of convolutional network error correction encoding which can correct a given error pattern.The core idea of this method is to equip a given error in the network to the source in where we design an error correction code that can correct the equivalent error.The advantage of this design is to avoid the problem of error spreading in the convolutional network which makes the encoding simpler.When decoding,according the conditions to decide whether in source or in sink.And then it adopts a proven method Viterbi decoding algorithm in classical convolutional codes.A new kind of network error correcting code(rank distance code)is studied in the second part of this master.Based on the application of rank distance codes to random network coding,we apply rank distance codes to convolutional network error correction coding and propose a scheme of error correction decoding for convolutional network based on Gabidulin code.We build a convolutional network error correction coding model firstly,and then under the model,we analyze the successful decoding conditions of the constant dimensional subspace code which is lifted by the rank distance code.Finally,we propose a Gabidulin code convolutional network error correction decoding scheme,the successful decoding condition of which is that the errors in the network satisfy 2wt(Br,0·Ei+Br,1·Ei-1+...+Br,τ·Ei-τ)≤2ε<dS(I(C))/2=dR(C).