Research On Combining Algorithms Of Differential Cooperative Communication Systems

Cooperative communications is one of the key technologies in 5G.The first advantage of cooperative communications is that it can not only obtain diversity gain but also is more suitable for the scenario with small terminals,such as Io T and sensor networks since it can be seen as a distributed MIMO.The second advantage of cooperative communications is that by introducing relay nodes and exploiting the non-linear characteristic of path loss,systems can obtain a huge gain of receiving power by reducing pass loss.Channel state information(CSI)can be obtained by channel estimation schemes.However,the computing complexity is proportional to the number of relays.As the fading of channels goes fast,the frequency of channel estimation goes higher as well,which further reduces the effective transmission rate and the spectrum efficiency.Also,the possible error in the estimated CSI will introduce additional degradation on the system performance.With the ability to bypass the high overhead and inaccurate channel estimation,differential modulation has been a hot topic for several years.By using diversity combining algorithm to combine the multiple received signal carrying same information,the overall SNR can be enhanced which can be used to improve quality of service,spectral efficiency and coverage rate.The most common combining algorithms includes maximal ratio combining and maximum likelihood combining,all need full CSI or partial CSI.However,in differential cooperative communications systems,nodes can not obtain CSIs,therefore those algorithms are not suitable for differential cooperative communications systems.Aiming at above problems,in this thesis,we study the combining schemes that not require CSI for differential cooperative systems.The detailed contents and aims of this thesis are as follows.Firstly,in DDF systems,the possible decoding error at node R can pass to node D by error propagation effect and therefore lower the error performance of the detector.In SC algorithm,the decision variables of direct and relay link have the same likelihood,therefore,SC can not lower the error propagation effect and hence can not achieve any diversity gain.To overcome this problem,we propose two WSC schemes for slow Rayleigh fading channels and analyze the performance.Different from SC scheme,both WSC schemes can obtain the diversity gain.The two WSC has different complexity and performance,therefore,they are suitable for complexity-limited and performance-limited scenarios,respectively.Secondly,energy harvesting technique is suitable for energy-limited communications systems and can prolong system lifetime,lower the cost and high energy efficiency.Cooperative communications systems using differential modulation do not need any CSI and omit the channel estimation and hence lower the energy consumption.Therefore differential modulation is suitable for energy-harvesting based cooperative communications systems.Although the SC scheme has a normal performance,it only needs process receive signals of one receive link.For these reasons,we first propose the SC scheme for energy-harvesting based DAF system which can achieve maximum energy saving in energy-limited systems.Also,we analyze the performance of the SC algorithm in slow fading Rayleigh channels and find the optimal power split factor for the relay node.Thirdly,linear combining(LC)algorithm require CSIs of all the links.Although LC adds demand of CSI,it can achieve better performance.It is known that Nakagami-m is a versatile model and can modeling outdoor and indoor mobile communication environment by adjusting the fading parameter m.Moreover,the distances of nodes in cooperative systems are also large,therefore,the links also have different channel characteristics.By modeling the links using Nakagami-m channels,we can simulate various practical scenarios and is more meaningful in project practice.For this reason,we study the performance of LC algorithm for SWIPT-based DAF systems over Nakagami-m fading channels,includes BER,diversity gain and asymptotic BER.Fourthly,as the speed of nodes increase,the channels become fast-fading channels.The precondition of the validity of differential modulation is that the channels do not change during the two successive usage,which is false in fast-fading channels.Therefore,differential cooperative systems experience a performance degradation over fast-fading channels.Until now,there has no effective combining algorithm for the DAF system over fast-fading channels.For this reason,we propose the maximum-likelihood(ML)combining algorithm for the DAF system over fast-fading channels.Although ML is the optimal algorithm,it requires CSIs of all the links which are not available in DAF system.Based on the ML algorithm,we propose sub-optimal ML algorithm,which requires only average channel gain of all the links.Finally,we derive the closed-form unified BER expression of the LC scheme using arbitrary weights over fast-fading channels.