Research On Power Optimization Algorithms In Massive MIMO Cognitive Radio Systems

The increasing demands for diverse ubiquitous communication services and smart connected devices have not only exacerbated the spectrum crisis,but also created enormous energy demand problems and environmental pollution problems.Thus energy efficiency(EE)becomes an important metric to evaluate the performance of wireless communications.The cognitive radio system(CRS)based on massive multiple-input multiple-output(MIMO)(i.e.,massive MIMO CRS),integrating the advantages of both CR and massive MIMO on spectral efficiency(SE)and EE,not only can achieve high degree of reuse of the licensed spectrum in space-timefrequency domain,but also can fully exploit the advantages of massive MIMO in the space dimension,to cater to the future explosive growth of traffic demands and alleviating the ensuing crisis.Power control is recognized as the key technology to explore the potential of massibe MIMO CRS.Therefore,that how to efficiently adjust the transmit power of CUs to obtain high EE subject to the constraints of primary communication and cognitive communications,has become one of the research focuses in the field of massive MIMO CRS.From the aspect of reducing the transmit power and maximizing the EE by optimizing the users’ transmit power,the power optimization algorithms in massive MIMO CRS have been systematically studied in this dissertation with considering the fairness among CUs and the impacts of channel estimation errors on the system EE performance.The main achievements and results of this dissertation are summarized as follows:1.In order to reduce the transmit power consumption of CUs,we proposed an efficient power control algorithm based on noncooperative game theory with a self-adaptive power threshold scheme.Considering the advantages of massive MIMO on the spatial multiplexing gain and reducing the uplink transmit power,we have firstly introduced the massive MIMO infrastructure into CRSs,and built a new system model,i.e.,massive MIMO CRS.Then,we have formulated the power control problems among CUs under this system into a noncooperative game problem and proposed an efficient power control algorithm to find the optimal solution of the above problem.Wherein we have given a self-adaptive power threshold scheme to adjust the power threshold according to the current communication information,in order to adapt wireless communication environment with constant changes.It has been verified that the unique Nash Equilibrium(NE)point exists and the proposed algorithm converges to this point.Simulation results demonstrate the effectiveness of the proposed algorithm.2.For the negative effect of circuit energy consumption of massive MIMO CRS on the system EE and the unfairness among CUs caused by the remained large-scale fading,we proposed an energy-efficient power control algorithm in the massive MIMO CRSs with perfect CSI,in order to maximize the cognitive network EE and guarantee the fairness of EE among CUs.Firstly,we have given a more realistic power consumption model to more accurately evaluate the EE performance for the massive MIMO CRS.Then,we have formulated the above problems as two nonlinear differentiable fractional programmings,i.e.,network EE optimization problem and fair EE optimization problem.However,both these problems are nonconvex and NP-hard,due to the fractional nature of EE and the inter-user interference.It is very difficult to find the optimal solutions in polynomial-time.To tackle this,based on the fractional programming and sequential convex approximation,we have proposed two efficient power control algorithms with an alternating iterative optimization scheme to find the optimal transmit power strategies.Simulation results verify the convergence property and effectiveness of the proposed algorithms.Besides,comparing to other methods,the proposed algorithms can achieve the same network EE with half of the base station antennas saved.3.Considering the channel estimation errors would seriously impair the fairness among CUs and the cognitive network EE,we proposed a globally optimal joint power iterative algorithm and a locally optimal joint power allocation algorithm.1)To ensure the robustness of the cognitive network and the fairness among CUs,we have formulated a fair EE maximization problem under the imperfect CSI case.However,it is very challenging to address the above problem,due to the fractional nature of fair EE,the interference caused by estimation errors,the inter-user interference,and the highly correlated optimization variables.With the aid of the gradient-based adaption method and the subgradient method,we have proposed an iterative power allocation algorithm based on the Lagrangian dual method to achieve the globally optimal transmit power strategy.Simulation results show that the proposed algorithm can achieve the best fair EE and good EE fairness among CUs.2)In order to improve the cognitive network EE performance under imperfect CSI case,we have modeled a net-work EE maximization problem.Similarly,it is very difficult to find the optimal solution in polynomial-time due to the nonconvexity and NP-hard nature of the above problem.Moreover,the iterative power allocation algorithm presented in part 1)does not work any longer for the sum rate item in the network EE.Based on the sequential convex approximation and concave-fractional programming,we have proposed a locally optimal power allocation algorithm,with achieving the suboptimal solution for the network EE maximization.Simulation results show that the proposed algorithm can achieve the best network EE performance and accommodate more CUs without impairing the performance of cognitive network.