Multicell Transmission Based On Random Beamforming And Interference Alignment

With the increasing popularity of smart personal devices as well as the rapid develop-ment of mobile wideband services,the global mobile traffic has been growing exponentially.Meanwhile,mobile devices and connections,including machine-to-machine type,are also increasing rapidly.To meet the demands of the booming mobile traffic and connections,the cellular networks are more and more heterogeneously and densely deployed,where dif-ferent types of access points coexist with each other,multiple heterogeneous cells overlap with each other,and users can access the network in a flexible way.The heterogeneous and dense deployment results in more complicated interference environments,which presents great challenges to interference management and data transmission strategies and restricts the improvement of cellular network capacity.Network operators should adopt appropri-ate interference management and data transmission strategies,which could accommodate to the characteristics of specific deployment scenario,interference and services,to expand the network capacity and to guarantee users' quality of services.Interfering multicell scenarios are elementary units in cellular networks for interference management and cooperative transmission.Due to the advantages of multiuser diversity and maximum degrees of freedom(DoF),random beamforming(RBF)and interference align-ment(IA)are applicable for downlink interfering multicell scenarios.Nevertheless,the application of the two techniques are still faced with many challenges.For example,the existing closed form sum rate analysis approaches for RBF are applicable only for limited scenarios,the closed form sum rate analysis approach for opportunistic interference align-ment(OIA)is still lacking,the existing closed form sum rate analysis approach based on IA' s limited feedback and feed-front is still deficient,and the strong interference feature in cellular interfering links has not been fully exploited.To cope with those challenges,we an-alyzed the performance of different transmission strategies in interfering multicell scenarios to determine their dominant regions.The main contributions of this dissertation are listed as follows.1.In view of the limited applicable scenarios of existing sum rate analysis approach,we propose a universal analysis approach for the closed form sum rate of multicell RBF.In the multicell MISO and MIMO interfering broadcast channels with heterogeneous users,it is proved that the achievable rate at low and medium SNR regime as well as the limiting rate at high SNR regime for each user can be approximated as a Gumbel type random variable,based on the SINR and SIR distribution features.Thus,the application scope of existing limiting throughput distribution(LTD)theorem based sum rate analysis approach can be expanded to the heterogeneous user scenario.After that,we continue to prove that in the multicell scenario with heterogeneous schedule probabilities each user' s achievable rate a-mong the whole SNR regime can be approximated to a Gumbel type random variable,based on the equivalent SIR and SINR distribution.Therefore,the existing LTD based approach is also applicable in the heterogeneous schedule probability case.The proposed general anal-ysis framework for the closed form sum rate of multi-cell random beamforming results in a concise enough format and a good scalability for different SINR/SIR distributions,homoge-neous and heterogeneous user deployments,the whole SNR regime,and homogeneous and heterogeneous schedule probabilities.The theoretical results resulting from the proposed analysis method have a good match with simulation results,which can be used for effective performance evaluation in multicell RBF.2.In view of the scarcity of effective sum rate analysis approach for OIA,the closed form sum rate analysis approach for interference subspace chordal distance based OIA is proposed,resulting in the accurate formula between sum rate performance and network parameters.At first,a general sum rate analysis framework is designed for OIA in MIMO IBC,based on the relationship between users' interference subspace alignment metric and the conditional achievable rate.On the basis of this framework,the closed form sum rate is derived for one kind of IBC with I base stations in which the sum of the number of data streams in each cell satisfies ?j=1I dj=N+1,where N is the number of receiver antennas at each user.Furthermore,the achievable rate for each individual user in such an IBC can be well approximated by the mean of a Gumbel distributed random variable,which greatly reduces the computation complexity.Finally,the DOF performance with the users' scaling law is confirmed with the obtained sum rate expression.Simulation results show that the proposed theoretical sum rate analysis approach for OIA has a good approximation quality.The proposed analysis approach herein is an effective performance evaluation method for multi-cell OIA for its concise format and low complexity.3.In view of the inefficiency of existing analysis methods for interference alignment's effective throughput,the closed form sum rate analysis approach for IA based on random vector quantization(RVQ)and scalar quantization(SQ)feedback is proposed,and the ac-curate expressions between the effective rate and quantization bits are established.At first,the RVQ based CSI feedback method is focused,the quantization error of RVQ is trans-formed into the equivalent quantization error from Gaussian channel error model,and the achievable rate under limited quantization bits based on RVQ feedback is derived.Secondly,the SQ based CSI feedback method is studied,which quantize and feedback each element of the channel matrices.The accurate relationship between the effective sum rate,quanti-zation bits and SQ's quantization error distribution is established.Finally the RVQ based decoding vector quantization and feed-front framework is considered,and the effective sum rate of IA under SQ feedback and RVQ feed-front is derived in an explicit expression as a function of quantization bits.The simulation results confirm that the theoretical results of the effective IA sum rates under different quantization methods are quite accurate,which can be used to determine the optimal quantization overhead in practical channel conditions.Besides,the theoretical and simulation results demonstrate SQ is much more convenient in CSI quantization for general IA scenarios than RVQ.4.In the interfering multicell scenario with unidirectional strong interference,the com-bined interference superposition and alignment scheme(CISA)is proposed.The designed CIS A scheme transmits two kinds of data streams simultaneously,i.e.,superposition streams and alignment streams.The former are handled via successive interference cancellation at those strongly interfered receivers,and the latter are aligned into receivers' interference sub-spaces.The joint transceiver designs for CISA is modeled as a weighted sum rate(WSR)problem,and the optimality equivalence condition between WSR and its corresponding e-quivalent weighted mean square error problem(WMMSE)is derived.According to the optimality equivalence,the alternate optimization approach for the local optimum of CISA is designed.For further rate improvement on the superposition streams,the combined in-terference superposition and selective alignment scheme(CISSA)is proposed,in which the superposition streams are allowed to be aligned into the interference subspaces of some se-lective users.The joint transceiver designs for CISSA is modeled as a WSR problem,then transformed into the equivalent WMMSE problem,and finally alternately optimized for a local optimum according to the optimality equivalence condition.Simulation results have confirmed that CISA and CISSA both improve the sum rate performance at low to medium regime,and keep the optimal degrees of freedom at high SNR regime.