Delay Performance Analysis And Optimization In Energy Harvesting Sensor Nodes

As an important part of the interconnection of everything,wireless sensor nodes will be widely deployed.However,the power supply of sensor node is a key factor restricting its application.Energy harvesting can use natural energy to power sensor nodes and lengthen the lifetime of nodes.One of the core problems of energy harvesting sensor nodes is how to sleep properly so that the nodes can have sufficient energy when there is data to send and achieve low delay.The researches on the delay performance of energy harvesting sensor nodes are not comprehensive now,and the researches on how to use the energy reasonably to avoid the deterioration of delay causing by the wake-up energy consumption in sensor nodes are very rare.For this reason,this paper have a study of the delay performance in energy harvesting sensor nodes.First of all,this paper studies the delay performance of the ideal wakeup nodes.Ideal wake-up nodes means awakening sensor nodes do not cost energy.In this paper we use a two-state Markov modulated on-off flow model to describe the burstiness of data packets arrival process and energy arrival process.And then we use a virtual queue to describe the coupling relationship between the occupancy of data cache and the energy in battery and obtain a set of differential equations.By solving these equations,the expression of average delay is obtained.And we explain how to design the battery capacity when a delay target needs to be satisfied.Then,we further discuss the problem of energy utilization in non-ideal nodes.Non-ideal nodes means awakening the node needs energy consumption.This paper presents an energy management strategy.In this strategy,when the data in cache or the energy in battery is cleared,the system goes to sleep;when the data length in cache and the energy in battery exceed an arranged threshold,the system will be awakened immediately and start to transmit data.In order to evaluating the delay performance of this strategy,a three-dimensional Markov chain is established to analyze the strategy.The analysis and simulation results show that there is always an optimal threshold when the battery capacity is finite,so that the node can get the lowest average delay.We further simulate the delay performance of the strategy when the data and energy arrival process is burst in the actual situation,and the results confirm the feasibility of the strategy.