Study On Interference Alignment For Multi-user Wireless Communication System

Interference is inevitable in wireless communication systems. Especially inmulti-user communication systems, multi-user interference severely limits thesum-capacity. When using conventional interference management approaches, such astreating-interference-as-noise, interference decoding and orthogonalizing, the spectrumefficiency of each user decreases as the user number increases, and the sum-capacity isinterference limited. Interference alignment is a novel interference managementapproach using the idea of overlapping interference space. Research shows that thesum-capacity increases proportionally with user number when using interferencealignment. Interference alignment has created a new field for improving sum-capacitywith limited resources, and becomes a research hotpot in wireless communication.To improve the diversity gain of blind interference alignment (BIA), this thesisproposes a new scheme which combines interference alignment with Alamouti coding.The proposed scheme achieves a diversity gain of2compared with that of1in theoriginal BIA scheme, thereby obtaining better BER performance at the same data rate.Based on the Alamouti structure, partial interference cancellation (PIC) group decodingis used at receivers, which reduces the complexity and further improves the BERperformance. In addition, the proposed scheme takes use of the idea of symbolextension. Because symbol extension is invalid when channel is relatively constant, thisthesis proposes to use reconfigurable antennas at receivers to solve this problem.Simulation results show that the proposed scheme obtains an8dB BER performanceimprovement in high SNR region at the same data rate as the original BIA schemeTo improve the system capacity of distributed interference alignment (DIA), thisthesis analyses the impact of different iteration algorithms on DIA performance,including minimum remaining interference algorithm and max-SINR algorithm. Basedon this analysis, the max-SNR interference alignment algorithm is accomplished whichcombines interference alignment and matched filter. Specifically, some degrees offreedom are used to achieve interference alignment, when the remaining ones areexplored to maximize the power of desired signals. Simulation results show that theperformance of max-SINR interference alignment algorithm is close to that of themax-SINR algorithm in most cases, while the complexity is lower. This thesis focuses on the impact of interference alignment on the systemperformance, and analyses the trade-off between diversity gain, multiplexing gain andcomplexity systematically. The study of this thesis establishes a theoretical foundationfor the implement of interference alignment in practical communication systems.