Performance Analysis And Resource Allocation Of Massive MIMO Based Heterogeneous Networks

Massive MIMO is one of the key physical layer techniques of 5G.When the base station(BS)is deployed with very large number of antennas,it can significantly enhance the system spectral and energy efficiencies by employing the low-complexity linear precoding methods.On the other hand,the hetero-geneous networks(HetNets)can improve the network capacity and coverage through deploying the different types of low power nodes in the coverage of macro BSs.Therefore,it is of great significance to take the advantages both of massive MIMO and HetNets to further improve the network performance.However,there are still many challenges to be addressed for the HetNets with massive MIMO.On one hand,with the dense deployment of BSs,the seri-ous interference and limited backhauls will become the main bottleneck of the performance improvement of HetNets.On the other hand,with the increasing number of antennas,the circuit power consumption of massive MIMO systems will be large.To solve the above mentioned issues,this dissertation will focus on the fundamental theory and technologies of massive MIMO based HetNets.The main works and contributions can be summarized as follows.1.The performance analysis and spatial DoF allocation for massive MIMO based HetNetsConsidering the serious cross-layer interference between macro and small cells in the massive MIMO based HetNets,the novel interference coordination user selection algorithm and spatial degrees of freedom(DoF)allocation al-gorithms based on the second-order statistics of channels are proposed.First,by using the random matrix theory,the tight low bounds of ergodic achievable rates for the macro and pico users are derived.Second,to reduce the overhead of the backhaul links,the interference coordination user selection algorithm based on the second order statistics of channels is proposed.The simulation results show the proposed algorithm can nearly achieve the sanme performance of the one based on the instantaneous channel state information(CSI).Further-more,considering the suffered inter-interference of macro users caused by the dense deployment of small cells,the joint spatial DoF allocation and user se-lection algorithm is proposed.The simulation results show that the proposed algorithm can strike a balance between the spatial multiplexing gain and inter-ference coordination gain of massive MIMO while satisfying the transmission rate constraints of macro users.2.The performance analysis and resource optimization of massive MIMO based HetNets with full-duplex wireless backhaulsTo address the challenges caused by the limited backhauls in the hetero-geneous dense networks,the scheme of wireless backhauls based on the full-duplex technology is designed.First,to migrate the multi-user interference and intercell interference,the zero forcing precoding scheme based on the projec-tion technology is proposed.Then,the ergodic achievable rate of macro and small cell users as well as backhaul links are derived by using the random ma-trix theory.To optimize the network performance with the limited capacity of backhauls,considering the fairness among the users,the joint user associ-ation and resource allocation problem is formulated to maximize the network logarithmic utility function.To solve this non-convex problem,the distributed iterative algorithm based on the Lagrange dual decomposition method is pro-posed.The simulation results show that the proposed algorithm can improve the network performance at least by 30%with a good convergence rate.3.The performance analysis and power allocation for the massive MIMO relay systems with imperfect ADCThe circuit power consumption of massive MIMO systems is large.To reduce the hardware cost of radio frequency chains,we focus on the massive MIMO relay systems with the low-resolution analog-to-digital converters(ADCs).Considering the imperfect CSI,the tight closed form of the user ergodic achiev-able rate is derived.The asymptotic analysis shows that the impact of the imper-fect ADCs on the data transmission can disappear with the increasing number of relay antennas.Second,to compensation the performance loss caused by the imperfect quantization,the joint user and relay power optimization prob-lem,which is non-convex,is formulated to maximize the system sum rate.To solve this problem,an optimal power allocation iterative algorithm based on the successive convex approximation(SCA)is proposed.Furthermore,to strike a tradeoff between the transmit power and hardware cost,the power allocation algorithm to minimize the total power consumption is also proposed.The sim-ulation results show that the effectiveness and convergence of the proposed algorithms.4.The performance analysis of the non-uniformly deployed mm Wave massive MIMO based HetNetsTo solve the problem caused by the non-uniform deployment of mm Wave massive MIMO based HetNets,based on stochastic geometry,the expressions of the associate probability,coverage probability as well as average rate are derived.Then,the impacts of the network parameters,such as the density of mm Wave small cells,the number of BSs in the cluster and the radius of the cluster,on the network performance are analyzed.The simulation results show that deploying more mm Wave small cells can enhance the coverage probability and average rate effectively.The increasing number of BSs in one cluster will lead to the decrease of coverage probability.But this impact can be reduced by deploying more mm Wave BSs.Furthermore,based on the theoretical results,by using the Jensen Inequality,the distance-based user association method is proposed to judge that which pattern of BSs can be associated by the users to achieve the best coverage probability.