Research Of Key Technologies Of Data Collection In Low-Duty-Cycle Wireless Sensor Networks

Low-duty-cycle wireless sensor networks(LDC-WSN)are new distributed self-organizing networks appeared in recent years,which are composed of large amounts of nodes with limited capacities(including communication,storage space and computation,and so on).LDC-WSN are often deployed in harsh environments that human beings are difficult to access to perform long-term tasks such as surveillance or tracking.They have broad application prospects in some fields such as national defense,industry and agriculture,etc.Therefore,they have significant economic and social benefits.Different with traditional wireless sensor networks(WSN),the nodes in LDC-WSN would sleep in most of their working period,and only wake up to perform sensing and communication tasks in a few time slots.Data collection is an important job of LDC-WSN,and its goal is to send the sensed data of nodes to the user for further processing.However,most of the nodes in the network are in sleep status in anytime,so the data is hard to be sent in real-time.As a result,the latency of data collection is large.Moreover,since LDC-WSN are deployed in harsh environments,communications are unreliable and the nodes are prone to be destroyed by outer forces.The unreliable communications would cause large amount of data retransmission,which would cost the nodes large amount of energy and time.On the other hand,the destruction of some nodes would cause the loss of their data.Therefore,how to implement quick,reliable and energy-efficient data collection in LDC-WSN is a challenging issue.This thesis researches key technologies of data collection in LDC-WSN,and its goal is to develop feasible data collection schemes with high robustness,good energy efficiency and low latency.The main research contents of the thesis are as follows:(1)Fast data preservation mechanism in LDC-WSN without fixed Sink.The nodes in LDC-WSN would sleep for most of their working period,which can largely reduce their energy consumption and extend the network lifetime.However,the sleeping of nodes would cause large communication latency since a sender node has to wait for the wakeup of its receiver.In some situations,there is difficult to deploy a fixed Sink in LDC-WSN since the environments are harsh.In such a network,the source nodes that sensed data should disseminate their data to other nodes in the network for redundant storage as soon as possible.Obviously,it is contradict with the feature that LDC-WSN have large latency.Existing data storage mechanisms did not consider the sleeping feature of LDC-WSN,so they cannot be applied in LDC-WSN.In this thesis,a distributed and fast data preservation mechanism(FDP)is proposed.FDP not only can perform data dissemination with low latency,but also enable the nodes efficiently store all of the data by using network coding in their limited storage spaces.After the data is stored,a mobile Sink can enter the network at anytime and anywhere,and it can recover all the source data when it visits parts of the nodes that remain survival.Theoretical analyses and simulations show that FDP can not only complete the process of data preservation with shorter time,but also cost lesser energy consumption.(2)Low-latency data persistence improvement algorithm in mobile LDC-WSN.In mobile LDC-WSN,the nodes would move from time to time,which makes the topology of the network change from time to time.Moreover,storage spaces of the nodes are limited.Therefore,it is a challenge that the data can be disseminated quickly and stored in the nodes.In this thesis,a novel data persistence improvement algorithm named LT-MDS is proposed.The algorithm uses a new infectious data dissemination method to send data to moving nodes.The data can be received by most of the nodes with low latency,which improves the reliability of the network.When a node receives a data,it would use LT-codes(Luby transform codes)to process the data and then store the encoded data.By this way,a node with limited storage space can preserve more data information.Theoretical analyses and simulations show that LT-MDS can complete the data dissemination and storage with low latency,and its data persistence is high.(3)Distributed circular pipeline work scheduling in low-duty-cycle wireless sensor networks.Long time sleeping would cause the increase of data query latency in LDC-WSN.How to schedule wakeup time of the nodes so that the latency can be minimized is a combinatorial optimization problem that is difficult to solve.A distributed circular pipeline scheduling algorithm is proposed to select nodes with high energy level to construct a path with minimized length between the start node and the target node in a query.In the path,nodes would wake up one by one likes a pipeline,which would benefit a node to send its data to its next hop neighbor without waiting for a long time.Theoretical analyses and simulations show that the algorithm can not only achieve lower latency,but also longer network lifetime.(4)Distributed multi-pipeline scheduling with low collision in LDC-WSN with unreliable communications.LDC-WSN can be deployed in harsh environments to perform long-term monitoring tasks,and they have broad application prospect.However,different with traditional WSN,the nodes in LDC-WSN would sleep for most of their working period,which causes large latency in data collection.In order to decrease the latency,existing works use multi-pipeline scheduling to adjust wakeup time of the nodes.By this way,nodes in multiple paths would wake up one by one,and a node can send it data to a receiver with high link quality.However,the works didn't consider the factors of communication collision and energy balance,i.e.,multiple sender node may select a same neighbor with high link quality as their receiver,which would cause serious communication collision and as a result cost the nodes high energy consumption for large amounts of retransmission.To solve the problem,a distributed low collision multi-pipeline scheduling algorithm is proposed.The algorithm considers some factors such as link quality,node's energy levels,number of a receiver's senders,etc.,to avoid the happen of the situation that many nodes select a same neighbor as their receiver.Theoretical analyses and simulations show that the algorithm can achieve lower latency and longer network lifetime.In general,this thesis researches some key technologies of data collection in LDC-WSN,and several algorithms with lower latency,higher energy efficiency and reliability are proposed.Therefore,the research has better application values and theoretical significance.