Research On Power Allocation Algorithm Of Massive MIMO System With Energy Harvesting

As one of the key technologies of 5G,Massive MIMO can significantly increases system capacity and spectrum efficiency by configuring a very large antenna array at the base station.However,with the increasing number of base stations and antenna size,the total energy consumption and carbon emissions are also increasing.Combining energy harvesting technology,developing green communication has become an inevitable trend of mobile communication development.In this paper,the energy harvesting technolog.y is applied to the Massive MIMO system,and the base station is powered by a hybrid energysupply method combining the collected energy and the power grid.The system capacity and power consumption model of Massive MIMO downlink is analyzed.Under the premise of considering the battery capacity limitation,An optimization problem with the objective of maximizing system energy efficiency and the variables of transmitting power and circuit power is established.And power allocation algorithms in two different cases are proposed,as follows:(1)Offline situation:Modeling the energy efficiency maximization problem as a fractional programming problem on the premise that the complete channel state information and energy harvesting process are known,and combining Dinkelbach algorithm and Lagrange dual method,an iterative offline power allocation algorithm is proposed to solve the optimal power allocation scheme.The simulation results show that the algorithm has a faster convergence speed and the system energy efficiency is improved by 4bit/J compared with the average power allocation algorithm.(2)Online situation:Based on the known CSI and energy collection statistical characteristics,two online power allocation algorithms are proposed in this paper,which are more realistic:online power allocation algorithm based on Markov chain and timing detection.the former model the energy arrival process and the channel change process as a Markov chain model,according to the information of the current epoch and transition probability,the base station can obtain a global situation over a period of time,then according to the offline algorithm for power allocation.The latter according to the current state information of the system and combining Dinkelbach algorithm and the Lagrange dual method to maximizes the energy efficiency in a single epoch,thereby maximizing the energy efficiency of the whole system.The simulation results show that the online algorithm attains a performance pretty close to the offline one,and it is much more feasible.