Modelling,Performance Analysis And Optimization For EH-WSN Node With Temporal Death

The limitation of energy supply in a wireless sensor network(WSN)is still a critical issue,and one of the most attractive approaches to extend the lifetime of wireless sensor networks is energy harvesting(EH)technology.However,the randomness and instability nature of the ambient energy source,jointly with the uncertain communication process give rise to many new challenges to the theoretical and actual application research of the energy harvesting wireless sensor network(EH-WSN).This dissertation investigates EH-WSN nodes and develops a stochastic model system to analytically analyze the performance metrics of EH-WSN nodes.These analytical models are further utilized to guide the design of EH-WSN,for example,determine the optimal energy buffer size and the optimal transmission policy.The main contributions of this dissertation are summarized as follows:(1)Modeling of the EH-WSN nodes with temporal deathIn this dissertation,a stochastic model system is proposed to characterize the EH-WSN node with temporal death.The model system consists of three stochastic processes:the energy harvesting process,the sensor operational modes and the residual energy process.Specially:·The energy harvesting processWe use a two-state Markov energy harvesting model to model the energy harvesting process herein,while extension to multiple-state model can be made with ease.·The sensor operational modesThe sensor nodes we considered have three operational modes:Sleeping,Listening and Transmitting.We use a M/PH/1 K queue model to capture the Transmission behavior.According to the construction of PH distribution,this model can be generalized easily.We can construct a more realistic PH distribution for the special scenarios when such as the MAC protocol,the transmission error control protocol,routing protocol,and the data buffer capacity size are given in the realistic application.·The residual energy processInstead of using a discrete state process to model the residual energy process,we establish a continuous fluid process to capture the residual energy in the EH-WSN node.Particularly,when the residual energy level is positive,the evolution of the node is in a normal manner,once the node runs out of the energy,the node then becomes inactive,losing its sensing and communication functionalities,and we refer to this special state as the temporal dead.(2)Performance metrics of the EH-WSN nodesLet the residual energy process in the EH-WSN node be a fluid level process,and let the energy harvesting process and sensor behavior process be a background process,we transform the model system of EH-WSN node into a modified Markov Fluid Queue(MFQ).Using the MFQ theory,we study various performance aspects of the EH-WSN node with temporal death,including the death occurrence probability,the residual energy probability density,the average queue length in the data buffer,the packet blocking probability and so on.In order to obtain the dropping probability of a given packet in EH-WSN node,based on the structure of the MFQ,we develop an auxiliary MFQ and derive the formulations of two types of the packet dropping probabilities,i.e.,the packet dropping probability due to energy depletion and that due to channel errors.Numerical examples are provided to illustrate the theoretical findings,and new insights into understanding the impacts of the parameters on the performance metrics are presented.(3)The designs of the EH-WSN nodesBased on the performance analysis of the EH-WSN nodes,we develop the system design for determining the optimal energy buffer size and the optimal transmission policy.To analyze the energy buffer size of an EH-WSN node,using the Finite Markov Fluid Queue(FMFQ),we propose an algorithm for obtaining the optimal energy buffer size by calculating the maximum reward,while the energy depletion and energy overflow probabilities are maintained below the given thresholds.We validate our theoretical results using a numerical example and further study the impact of the parameters.·In order to model and analyses the transmission nature of the EH-WSN node,based on the theory of Mutil-Layer Markov Fluid Queue(MLMFQ),we propose an algorithm for calculating the optimal transmission policy of the EH-WSN node.Finally,we validate our results using a numerical example and report some findings from our numerical studies.The frameworks proposed in this dissertation are extensible,and the results obtained herein are of practical and theoretical importance in the design,development and optimization of EH-WSNs.