Research On Cross-layer Optimization Strategy In Energy-harvesting Wireless Sensor Networks

The problem of energy constraints in sensor nodes is a main bottleneck that hinders development of Internet of Things.Even if the 5th-Generation mobile communication technology(5G)can provide Internet of Things with low-power and massive access points.Wireless Sensor Networks(WSNs)still face the problem of energy constraints in sensor nodes.To solve this problem,Energy-Harvesting Wireless Sensor Networks(EH-WSNs)came into existence.In EH-WSNs,sensor nodes with the function of harvesting energy can supplement energy for themselves in time without replacing batteries.EH-WSNs are not only prolong the lifetime of networks,but also achieve the purpose of protecting environment.However,with the influence of external environment and nonideal characteristics of sensor nodes' hardware,the energyharvesting process is random and uncertain.The problem of using harvested energy to complete data transmission efficiently becomes an urgent issue to resolve in EH-WSNs.At present,researches on the energy usage and data transmission in EH-WSNs are usually based on one certain layer of the network protocol stacks,while the actual process of data transmission is related to several layers of the network protocol stacks.Cross-layer optimization in EH-WSNs is expected to improve network performance further.In this thesis,the cross-layer optimization strategies in EH-WSNs are studied and several methods are as follows:(1)A cross-layer optimization strategy of joint routing and link scheduling is proposed in this thesis.The optimization strategy is based on network layer and data link layer to establish a mathematical model.At the same time,nonideal characteristics of batteries are considered.Then,the cross-layer optimization problem is transformed into a Mixed Integer Linear Programming(MILP)problem.On the basis of modeling,a cross-layer optimization strategy that runs at sink node is put forward.By means of altering the values of nonideal characteristics parameters of batteries,the simulation experiments verify that the effect of different nonideal characteristics on transmission time.The final simulation results show that both charging/discharging efficiency and energy leakage ratio have certain effects on transmission time,while battery capacity has no effect on transmission time.For different energy storage structures,HarvestUse-Store(HUS)energy storage structure is superior to Harvest-Store-Use(HSU)energy storage structure.(2)A cross-layer optimization strategy of joint routing and power control is proposed in this thesis.The optimization strategy is based on network layer and physical layer to build a mathematical model.Then,the cross-layer optimization problem is transformed into a MILP problem.Based on modeling,a Centralized Cross-layer Optimization Strategy(CCOS)is proposed,which is operated at sink node.Owing to limitations of CCOS,a Distributed Cross-layer Optimization Strategy(DCOS)was raised,which renders each sensor node route and control its power according to state information of its neighbor nodes.The simulation results show that CCOS perform better than DCOS in small scale networks,while DCOS has better performance in large scale networks.At last,the simulation verifies that DCOS with power control can improve the performance further.In conclusion,the problem of cross-layer optimization in EH-WSNs is investigated in the thesis.Cross-layer optimization strategies of joint routing and link scheduling as well as joint routing and power control are put forward respectively.Joint routing and power control of cross-layer optimization strategy includes CCOS and DCOS.The strategies in this thesis improve data transmission efficiency and are conducive to promoting further development and maturity of the Internet of Things industries in 5G era.