Research On Interference Alignment In Asymmetric Two-way Relay Networks

Interference alignment, a novel interference management method, introduces theconcept of the degrees of freedom. The technique makes the interference signals sent bydifferent senders overlapped on a limited dimensional space provided by the receivers.Accordingly, the dimension of the interference signal subspace and the desired signalsubspace can be compressed and increased, respectively, which makes it possible toenlarge the degrees of freedom of the system. Interference alignment is very useful forthe reliable transmission of the large data in the multi-user interference for the nextgeneration wireless communication system. Besides, it plays a major role in theresearch of the interference network capacity.Compared with traditional one way relay networks, the two-way relay X networksoffers much higher spectrum efficiency, however, leads to more complex interference.In this thesis, the asymmetric two-way relay network is used as the model to researchthe interference management scheme and the feasibility conditions, and thus to analyzethe degrees of freedom of network and the capacity. The main work and innovations ofthis thesis includes:For the asymmetric two-way relay interference networks, an interference alignmentscheme based on the maximum signal to interference noise ratio (Max-SINR) criterionis proposed, which is more suitable for the low or the intermediate transmit signal tonoise ratio (SNR) scenarios. First, the user signal alignment scheme, relay forwardingmode and user interference suppression algorithm are proposed. Second, the feasibilityconditions of the proposed scheme in multiple access control (MAC) phase andbroadcast (BC) phase is analyzed. In addition, the achievable maximum degrees offreedom (DoF) of the network is derived. For the proposed scheme, the annoying signalprocessing in the relay nodes is avoided; the computational complexity of the user endsis mediate, and thus the overall computational complexity of the entire system issignificantly reduced. Furthermore, the proposed algorithm improves the networkcapacity, especially for the low or intermediate transmit SNR scenarios. Also, thetheoretical maximum value of the network degrees of freedom is yielded.For the asymmetric two-way relay X networks, the drawbacks of the existed zerospace criterion-based algorithms, the complexity and diversity of the X networks areconsidered simultaneously. With these in mind, the thesis proposed an interference alignment scheme, which considers both the desired signal and interference signalcomponents and uses the networks sum rate as the objective function. First, the usersignal alignment method and the feasibility conditions are proposed. Second, themathematical model for the network sum rate is obtained based on the spatialrelationship between user signal vectors, and the final objective function isapproximated by using a series of optimization. Finally, the relay pre-coder matrix isobtained as the linear programming algorithm. The proposed scheme significantlyimproves the network sum rate compared with existing relay processing scheme, whilesimplifies the complexity of the conventional X network.