Research On Energy Scheduling In The Rechargeable Internet Of Things

As the key component of the next generation information technology,Internet of Things(IoT)is entering the critical stage of "border-crossing integration,integrated innovation and scale devel-opment".However,the energy supplement problem has always been obstacle to further develop-ment of IoT.In recent years,rechargeable Internet of things integrates wireless charging,energy management,data sensing,communication and calculation,and becomes a reliable solution to en-ergy supplement of IoT.Due to its potential in smart home,intelligent warehousing,intelligent transportation,and many other areas,rechargeable IoT has attrated considerable attentation from both academia and industryRechargeable IoT harvests energy from surrondings via multiple technologies,to free devices from the restiction of network lifetime and deployment cost due to energy storage,and enables de-vices to utilize energy effectivelly via energy management technology.Energy scheduling plays an important role in improving network utility.Thus,an amount of researchers have investigated on this area.However,there are still some problems to be solved,especially in stable wireless energy harvesting,data collection in ultra low-power conditions,decoupling of charging process and dis-charging process,etc.This dissertation focuses on energy scheduling optimization in rechargeable Internet of Things.Effective energy scheduling schemes are designed to improve network utility from the aspect of network deployment,energy replenshment and energy assignment.The main work and contributions are summarized as follows:A brief review on the background,applications and related works about the rechargeable Internet of Things is provided.The relative research challenges are also sumarrized2.Research on the co-deployment problem of chargers and sink stations in the rechargeable In-ternet of Things.Existing commertial rechargeable IoT facilities design the charger and the sink station separately.This dissertation considers the co-deployment problem of chargers and sink stations due to this design.This dissertation investigates an algorithm to minimize network deployment cost while preserving the communacation demand.The primal prob-lem is divided into two subproblems,which can be transformed to the max-flow model.An approximation algorithm with ln R/ζ worst-case bound is proposed to solve each subproblem,respectively.Further,the solution to the primal problem is optimized by solving two sub-problems alternatively to achieve via extensive simulations.3.Research on the wireless charging scheduling problem in dockless bike-sharing systems.In recent years,dockless bike-sharing has become a popular application scenario of IoT.How to provide robust energy supplement for the smart lock module,is the key to improving user experience and reducing management cost.This dissertation introduces RF wireless charging technology as a solution.An RF wireless charging sensor node is designed to be integrated on a bike’s basket,such that the mutual interference during charging process and space occupation can be reduced.A greedy-based single charger scheduling algorithm is designed to reduce charging delay.The algorithm is extended to multiple charger scheduling for large-scale scenarios by applying dynamic programming.The whole system has been successfully implemented on a dockless bike-sharing system,which proves its effectiveness.4.Research on the operation state scheduling problem in the rechargeable IoT.Existing com-mertial rechargeable IoT sensor nodes can not work and harvest energy simultaneously.This leads to a new design challenge of optimally scheduling sensor nodes’ operation states:work-ing or recharging,to achieve a desirable network utility.This dissertation investigates this problem.In the single-hop case of small-scale networks,the operation state scheduling prob-lem is transformed to a linear programming problem,and an optimal analytical solution is ob-tained.In the multi-hop case of large-scale networks,this dissertation analyzes the spatiotem-poral coupling caused by time-varying network topology.A scheduling algorithm based on Lyapunov optimization is propsed to decouple the problem and achieves a MN/2V worst-case bound.Field experiments on Powercast platform and numerous simulations demonstrate the effectiveness of our design.5.Research on the battery-lifetime-aware scheduling problem for electric bus fleets.Electric bus network is another important application of rechargeable IoT.However,the high man-ufacturing cost of battery packs and limited battery lifetime hinder its further development This dissertation considers the electic bus scheduling problem from the view of battery life-time.This part intends to optimize the departure intevals to minimize battery degradation rate while preserving user experience.Leveraging practical bus GPS and transaction datasets,de-tailed analysis of passenger behaviors and a reliable prediction model for passenger arrival rate at each station are provided.Then,Lyapunov optimization is applied to obtain an elec-tric bus scheduling strategy with reliable performance guarantee on both battery degradation rate and passengers’ service quality.Experimental results on practical bus operation datasets verify the effectiveness of our system.6.The conclusions are drawn with future work at the end of this dissertation.