A Research On Transmission Over Relay With Network Coding

With the rapid development of the economy and society,users' demands for large data volume and low latency communication are increasing,which is a severe challenge for the current network.In modern communication technology,especially with the advancement of 5G technology,wireless communication is developing in the direction of high speed and high bandwidth.The relay technology can achieve higher link capacity and better coverage,and plays an important role in the 5G transmission scenarios.The network coding technology is used to encode and process the received data packets at the relay node,which can improve the throughput of the network transmission.At the same time,the fountain erasure-correction feature of the network coding can also reduce the feedback delay and simplify the communication protocol.However,in practical applications,the relay buffer space may limit the application value of network coding technology.It is one of the hotspots of current research to optimize network coding transmission efficiency under the condition of finite-buffer relay transmission.This paper focuses on the application of network coding around finite-buffer relay,and aiming to improve the transmission performance of the relay link.The main work and achievements are as follows:(1)Summarize the advantages and disadvantages of existing network coding techniques such as random linear network coding and sparse network coding.Study the impact of these technologies on transmission performance in a relay link communication scenario.(2)For the finite-buffer relay link scenario in which a random linear network coding is applied,the relationship between the transmission coding coefficient overhead,the transmission completion time and the finite field size is studied.In the two-hop wireless transmission link of burst packet loss,based on the Gilbert-Elliott channel model,the absorbing Markov chain is used to model the process of receiving the innovative data packets of the relay node and the destination node and channel state transition,and analyze the transmission completion time.An objective function that maximizes the effective rate is established based on the transmission overhead and the completion time.By optimizing the finite field size algorithm,the optimal finite field size is determined to maximize the effective rate of transmission.(3)For the higher decoding complexity and the coding coefficient transmission overhead of random linear network coding,it is proposed to use fixed degree distribution batch sparse network coding method in finitebuffer relay transmission link,and fix the transmission quantity of each batch.The batch encoding transmits and fixes the batch transmission size,which can effectively reduce the coding coefficient transmission overhead.At the same time,the sparseness can be maintained after the relay node reencoding.At the receiving end,the batch sparse feature is fully utilized,and decoding is performed by using an overlap-aware decoder to reduce the decoding complexity.(4)For the relay transmission scenario with state feedback of nodes,the fixed batch transmission scheme can be further optimized.Two optimization methods based on Markov decision process and instantaneous sub-matrix optimization strategy are proposed,respectively.The batch transmission size is dynamically adjusted according to the feedback information of the receiving end.The Markov decision process strategy dynamically adjusts the number of transmission coded packets per batch,that is,the batch transmission size,by establishing a value function which maximize the effective "sum" rate of each batch transmission,which reduces the transmission completion time.The instantaneous sub-matrix strategy is based on each feedback information and establishes an objective function according to the current minimum completion time,and determines the optimal batch transmission size for each batch transmission,which reduces the overall transmission time and improves transmission efficiency.