A Research On Dynamic Resource Allocation Algorithm In Data And Energy Integrated Network

Data and Energy Integrated Networks(DEIN)is an emerging network,which is promoted by the concept of green energy in the current era.It is an extension of an earlier energy harvesting network.Traditional energy harvesting networks only focus on harvesting natural energy,which is difficult to be a reliable energy source since the uncertain of the natural energy reception.Sensors may harvest energy from RF signal broadcasted by the base station in DEIN,thus energy source becomes reliable and controllable.Different from the traditional wireless communication,the base station has a continuous energy supply while users have not and need to rely on energy harvesting to complete their own communication needs by using energy harvesting equipment.Since the energy harvested from rf signals is very low,how to allocate dynamic resource in DEIN to utilize energy efficiently has become an important research direction.The dynamic resource in DEIN mainly include time resource in time-domain,beamforming resource in space-domain,frequency resource in frequecy-domain and schedule resource in user-domain.This thesis first proposes a single-basestation multi-user multi-antenna DEIN model based on tranditional wireless powered communication networks(WPCN).This DEIN model extends WPCN model by setting the base station to full duplex mode,where the base station transmits energy signal in the DL transmission and receives the data signal transmited by users in the UL transmission at the same time,and the energy transmission time is greatly increased.This thesis eliminates self-interference noise produced by full duplex by designing both the beam formers at the transmiting antennas and the receiving combinations at the receiving antennas.Due to the full duplex,the throughput of the model is affected by the user order of the UL transmission,which should be taken into account as dynamic resource to optimize.This thesis maximize both the sum-throughput and the fair-throughput by optimizing the dynamic resource in time-,space-and user-domain.Since the multi-domain resource is optimized and the discreteness of the user scheduling,the mathematic models of both the sum-throughput and the fair-throughput maximization are proved to be non-convex problems.Therefore,algrothims based on both block descend and dynamic progamming are proposed to deal with the coupling and the descreteness in the optimizations.The simulation results reveal that the proposed algorthims converge to sub-optimal results in a fast speed,and the proposed model performs better than WPCN model when the number of the antennas is relatively lager.In order to solve the problem that WPCN can not transmit data in the UL transmission,this thesis proposes a single-basestatin multi-user multi-antenna DEIN model using SWIPT in the DL transmisison.Equiped with power splitting circuits,users split the power of the signal transmited by the base station into two parts by power splitting factors to complete both the energy and the data receiptions.Each user has its own DL throughput demand in this model.This thesis maximize both the sum-throughput and the fair-throughput by optimizing the time slots of the UL and DL transmission in the time-domain,the beam formers at the transmiting antennas in the base station in the space-domain,and the power splitting factors in the power-domain.The mathematic models of both the sum-throughput and the fair-throughput maximization are proved to be non-convex problems since there exists coupling variables.As a result,algorithms based on block descend are proposed to obtain a sub-optimal result.The simulation results reveal that the algorithms coverage to a sub-optimal result quickly,and the UL throughput remains the same as the WPCN system when the DL throughput demand is relatively small,which means that the proposes model is suitable for the current IOT scenario.