Rateless Coded Uplink Transmission Design With Partial Channel Information In C-RAN

With the rapid development of mobile communications,the diverse business needs,ultra-dense networking and ultra-high mobility support,as well as the complexity and inefficiency of interoperability caused by multi-network coexistence pose a great challenge to the current network architecture.Cloud radio access network?C-RAN?as a new wireless access network architecture provides the possibility to solve the above problems.Compared with the traditional wireless access network architecture,the channel state in C-RAN is more complex and the channel state information?CSI?is full of uncertainty,so the channel coding scheme in C-RAN has higher requirements.Rateless code has good adaptability.Even if there is no instant channel state information,the performance of the optimized rateless code is close to the theoretical limit.Therefore,rateless code can be applied to C-RAN channel coding in the physical layer.Based on the above discussion,this paper mainly studys rateless coded uplink transmission design for C-RAN of partial channel state informations,main contents and contributions are as follows:?1?The design and optimization of rateless coded uplink transmission scheme for single user scenario in block fading channel in C-RAN are studied.In the scheme,the user does not need instant CSI,but only with statistical channel information.Firstly,the user uses the LDPC code with high bit rate as the pre-encoded Raptor code to encode the information to be sent.Then,the codeword is modulated and sent to each remote radio head?RRH?covering the user,and the pre-processing unit of each RRH gets baseband signal after pre-processing with the received signal.Then,the baseband signal is passed through the scalar quantizer with uniform quantization interval at RRH,and the quantized signal is transmitted to the BBU pool through the high-speed link.Finally,the BBU pool uses the decoding algorithm based on belief propagation?BP?to jointly decompress and decode the received signal.When the BBU pool correctly recovers the original information of the user,the acknowledge?ACK?signal will be fed back to the user and the user will stop sending the information.Furthermore,BBU pool iterative decoding is analyzed based on external information transfer?EXIT?relationship.Then on the premise of successful decoding recover the user information and to minimize the average code length as the goal,the optimal user's rateless code degree distribution is obtained by solving the linear programming?LP?problem.By the simulation results,compared with the traditional binary erasure channel?BEC?degree,the optimized rateless code performs better on the average throughput of the system,about 13%off the theoretical limit.?2?The design and optimization of rateless coded uplink transmission scheme for multi-user scenarios in block fading channel in C-RAN are studied.Similar to the single-user case,each user has only statistical channel information.In the multi-user scenario,each user's coding scheme and RRH signal processing scheme are similar to the single-user scenario.In the BBU pool,information recovery of multiple users is realized through the joint decompression,multi-user detector and iterative structure of channel decoding based on belief propagation.Furthermore,the joint optimization of the degree distribution of rateless code for each user in the multi-user scenario is studied.Different from the single-user scenario,the goal of minimizing the average code length is converted into minimizing the required signal-to-noise ratio?SNR?when the average transmission rate of the system is met?i.e.the average transmission code length of users are fixed?,and then the optimal degree distribution is obtained by differential evolution algorithm?DE?.The simulation results show that the rateless code transmission scheme with optimized degree distribution is better than the BEC degree distribution in terms of bit error rate?BER?and average system throughput.In the 8-bit quantization scheme,the bit error rate of10^-4 is achieved with less than 1.14 overhead,and the throughput is only 11%less than the theoretical limit.