Study On Reliable Transmission For Cooperative Communication Systems Based On Network Coding

Multiple-Input-Multiple-Output (MIMO) technology plays a fundamental role to improvetransmission reliability and spectrum efficiency in wireless communication system. However, dueto size, hardware complexity and cost limitations, it is difficult to be deployed in engineeringpractice. Recently, as an effective way to achieve MIMO, Network Coding (NC) cooperativecommunication has attracted widespread research attention, for on the one hand, by forming virtualMIMO system, cooperative communication can obtain diversity gain with low cost, while on theother hand, the application of NC is efficient to make up the loss of spectral efficiency caused bycooperative communication. These above advantages also make NC cooperative communicationone of the key technologies for the subsequent evolution in next generation mobile communicationstandard (LTE-A). However, since NC is originally designed for the error-free wired channel, whenit is applied in error-prone wireless channel, the reliable transmission issue must be handled firstly.Therefore, this thesis focuses on the reliable transmission of NC cooperation communication, andthe main contributions are summarized as follows:1) In the research of error propagation mitigation based on NC in multiple access channels, anovel link adaptive smart NC algorithm is proposed based on Log Likelihood Ratio (LLR) toimprove the limited adaptation accuracy of instantaneous channel gain. In the proposed algorithm,LLR is employed as the reliability instruction, and the equivalent instantaneous channel gain issolved by setting the LLR-based Bit Error Rate (BER) analytical formula and the instantaneouschannel gain-based one as simultaneous equations. This equivalent parameter is then employed toreplace the instantaneous channel gain to operate dynamic power scaling. The results indicate thatthe proposed algorithm is effectifftive to mitigate error propagation and can achieve full diversityorder.2) In the research of user fairness of NC cooperative communication in multiple access channels,a fairness-oriented power allocation algorithm is proposed to handle the inefficiency of theconventional metric in reflecting the overall network reliability. In the proposed algorithm,simultaneous outage is exploited as the metric to evaluate user fairness, and to optimize it, the specific powerallocation constraints are derived. On the basis of that, the signals of users are superimposed in specificcalculated proportion at relay and then decoded by Maximal Ratio Combination (MRC) and SerialInterference Cancellation (SIC) at destination. The results show that the proposed scheme can supply approximately equal performance promotion for each cooperation-participated user andimprove the simultaneous outage behavior of the system, as well as the transmission reliability of thewhole cooperative group.3) In the research of NC in the multi-source multi-relay cooperative networks, a MaximumDistance Separable (MDS) NC is proposed to fully exploit the existing links between relays. In theproposed scheme, Relay Ordering (RO) algorithm is introduced for the accumulation of mutualinformation through MDS codes. The relay transmission order is scheduled according to the numberof outage links, and then the relays who initially fail to decode from sources are kept on decodingthe signals from other relays, in order to recover the NC opportunity. The outage and full diversityprobabilities are theoretically derived, furthermore, the scheme is proved to own the merit ofdiffusion effect. The results show the proposed scheme can substantially improve the outageperformance of the system.4) In the research of the decoding algorithm design and optimization for asynchronousPhysical-layer Network Coding (PNC), a novel phase equalization based on pre-distortion isproposed to mitigate the impact of the carrier phase offset. In the proposed scheme, the intentionalphase distortion is pre-brought into the symbols of one user, and then, it is employed by the relay toequalize the received signals and modify the updating equations in the belief propagation decodingprocess. Since the closed-form BER formula is difficult to be achieved due to the constellationdistortion, a phase gain-loss factor model is founded, and based on it, the upper bound of theequalization performance is theoretically proved, while the phase distortion vector and algorithmrobustness are also optimized. Moreover, a threshold is proposed to maintain the processing gainand avoid unnecessary distortion. Finally, to mitigate the accuracy fluctuation of phase estimation,the relative symbol offset is introduced into the maximal likelihood estimation as the weight factor.The results indicate that the proposed algorithm can substantially improve the BER performancewithout real-time feedback channel and ideal estimation assumption.