Research On Self-maintainable Topology Control Algorithms For Wireless Sensor Networks

Wireless sensor networks (WSNs) offer a flexible low-cost solution to the problem of informationcollection in the real world, especially in places with limited accessibility or that represent danger tohumans, which has indicated a large application foreground in the fields of national defense, industry andagriculture, environment monitoring, medicare, and so on. WSNs are made of resource constrained wirelessdevices with the purpose to run the assigned tasks as long as possible, which require energy efficientalgorithms and protocols to maximize the network lifetime. One of the algorithms is topology controlcomposed of topology construction and topology maintenance which can not only reduce energyconsumption and increase the network lifetime, but also provide a reliable and optimal topology structurefor the upper algorithms and applications. Therefore, it is of profound practical significance for furtherdevelopment of WSNs to study on the topology control techniques.This dissertation expands the knowledge of topology control in the following ways. Firstly, itintroduces a comprehensive taxonomy for topology control algorithms, and points out the limitation of thepresent classifications following an extended definition of topology maintenance, and designs a generalmodel of topology maintenance. In order to learn more about the essence of the topology maintenancetechniques, it presents a classification, comparison and analysis for the current topology maintenancetechniques, and discusses the existent deficiency and future trends. Secondly, it proposes two new powercontrol based self-maintainable topology control algorithms: LRTC and AFCS. These two algorithmsconstruct the network topology by the ways of open and closed power control, and trigger the topologymaintenance phase based on energy and nodes’ degree. Finally, it presents two self-maintainable clusteringtopology control algorithms: DMC and NFCT. Using lightweight clustering mechanisms, the twoalgorithms construct cluster topologies and trigger topology maintenance phase based on the deviation ofresidual energy and fault respectively. The results of simulations show that these four algorithms canimprove the energy efficiency and increase the network lifetime.Power control is an effective technique of topology control for wireless sensor networks, whosepurpose is to reduce energy consumption as well as keep the basic properties such as connectivity andcoverage of the network. LRTC consists of three phases, and in the first phase Compiling_of_Lists, theneighbor list of each node which contains items of the minimum transmission power needed to makeproper communication between the node and its one-hop nodes is built. Moreover, a link cost functionbased on LQI (Link Quality Indicator) and residual energy is defined in the second phaseConstruction_of_Toplogy. Then a connected topology is constructed by using the function to judge whetherthere is multi-hop path with every link cost less than that of any two direct communication nodes, and anytwo one-hop nodes in the topology use the minimum power to communicate each other. Finally, theMaintenance_of_Topology phase is triggered locally when a certain node’s residual energy is lower thanthe average residual energy of its one-hop neighbors. However, the actual control effect of LRTC is hardly to be acquired due to its open loop control. Therefore, AFCS introduces the closed loop control theory usedin automatic control field to adjust the transmission power based on node degree. A self-adjustable fuzzycontroller is designed with two inputs the node degree deviation and its ratio with power deviation in orderto form a closed loop control. The controller changes the transmission power till the node degree is in thepredefined error range. And during the network running, the topology maintenance phase is triggered whilethe current node degree is out of the error range, and the expected node degree is updated according to theresidual energy of the node and its neighbors, the last node degree, the maximum and minimum nodedegree.Clustering is another effective technique of topology control for WSNs, which reduces the networkenergy consumption by assigning different roles to nodes with different burdens. The DMC algorithmcalculates the clustering parameter of every node based on the residual energy of the node and its neighbors,the distances between the node and its neighbors in order to make it larger probability for the nodes withmore residual energy and more neighbors in low energy and short distance. And DMC uniformly distributesthe cluster heads by restricting the location of their neighbor cluster heads to form a connected topologystructure. Furthermore, a local topology maintenance phase is triggered when the residual energy of acluster head is lower than the average residual energy of the nodes in the same cluster. However, once acluster head turns to be faulty, then the information in the cluster even the neighbor clusters can’t betransmitted to the sink. That is why NFCT is proposed to deal with the fault in the clusters. At first, acluster head backup mechanism is used to improve the fault-tolerant ability of the clusters in NFCT, andespecially a lightweight real time fault detection method is adopted to detect faults in a cluster. Once a faultis detected, the topology maintenance phase is triggered to deal with the fault. Finally, the topologymaintenance phase is also triggered to turn the backup cluster head to be cluster head in order to make thetopology be stable and more fault tolerant when the residual energy of a cluster head is lower than theaverage residual energy of the nodes in the same cluster.This dissertation focuses on the research of self-maintainable topology control algorithms based onpower control and clustering, and discusses a serial of issues such as link cost, expected node degree, evenclustering, fault tolerant clustering, and trigger criteria for topology maintenance mechanisms (based onenergy, node degree, fault, etc.) in detail. So, the four algorithms proposed in this dissertation have a greatsignificance for reference to design topology control algorithms.