Studies On Coding And Decoding Methods In Multi-cell Massive MIMO Systems

Technologies of mobile communications have entered into a rapid development stage.High quality of communications and high rate of data transmission are required urgently,and the traditional Multiple-Input Multiple-Output(MIMO)technology cannot satisfy these requirements anymore.As an extension or expansion of the traditional MIMO technology,massive MIMO technology emerges recently.Its high spectrum efficiency and high power efficiency make it become one of the key technologies of the next generation mobile communication systems.Massive MIMO systems have been widely studied in recent years.And research results have shown that performances of the massive MIMO systems with dozens or even hundreds of antennas on the base station heavily depend on the accuracy of the Channel State Information(CSI).Generally,acquisition of the CSI is based on pilot-based channel estimation,and the accuracy of the CSI depends on selection of pilot signals.In fact,in order to ensure the accuracy of the CSI,the pilot signals of all users must be orthogonal.However,it will consume a lot of resources to satisfy the orthogonality.Considering the limitation of spectrum resources and channel coherent time,scheme of pilot reuse has been adopted usually,where users in different cells use the same pilot signals,while users in the same cell use orthogonal pilot signals.In the pilot reuse case,because the pilot signals of typical cell and interference cells cannot be distinguished accurately by the receiver,the accuracy of channel estimation will be reduced largely,and the performance of system will be degraded greatly.This phenomenon is called as pilot contamination.One of the most direct influences of the pilot contamination is existence of error floor.For uplink massive MIMO systems with pilot reuse,this thesis proposes coding and decoding schemes to mitigate the effect of pilot contamination.Specific research includes two aspects:on the one hand,by analyzing the cause of error floor based on a simple case,this thesis proposes a coding scheme to reduce the error floor.Details of implementation of the proposed scheme and several design examples are presented.On the other hand,because the coding based optimal decoding method is Maximum-Likelihood(ML)decoder,its decoding complexity is very high.To reduce the complexity,this thesis presents two decoding methods with lower decoding complexity,namely OPD(the Oblique Projection based Decoder)and SOPD(the Simplified Oblique Projection based Decoder).These two decoding methods make use of the principle of affine projection to eliminate the interference from the interfered users.Since the OPD and SOPD are all based on partial ML decoder,the decoding schemes can realize the balance of the performance and the decoding complexity of the system.Through the analysis of simulation results,it can be seen that the proposed coding scheme can alleviate the effects of the pilot contamination and reduce the error floor effectively.Also,the performance of OPD and SOPD proposed in this thesis is much better than the MMSE decoder.Furthermore,with the increase of the number of antennas and users in each cell,the performance of OPD and SOPD will be improved constantly.