Research On Topology Control In Wireless Sensor Networks

In recent years, extensive efforts have been lucubrated on wireless sensor networks (WSNs). Because of their particular features and wide potential applications, WSNs have become a new attractive IT technology. Their essential features include: they require no or a few infrastructure, the nodes carry limited energy, and the wireless communication can be easily interrupted. So the basic tasks are to guarantee network connectivity, to improve the energy efficiency, and to assure network fault-tolerance in WSNs. There are still many problems to be resolved for further performance improvement.In WSNs, topology control (TC) is one of the most effective methods to ensure network connectivity and to enhance the reliability and fault-tolerance of network connection. The thesis presents in-depth study on TC and its cross-layer impact in order to improve energy efficiency and enhance network connectivity and fault-tolerance by means of TC. The main contributions of the thesis are as follows:(1) The prior works on TC are investigated and the existing TC algorithms are analyzed regarding their complexity, stabilization, energy efficiency and fault-tolerance. The thesis figures out the unsolved problems on TC.(2) One of the basic ideas in Control Theory, the closed-loop control, is introduced into TC, and two TC algorithms based on closed-loop, PIDTC/FCTP, are designed. The algorithms require less convergence time and are more stable than the existing relative algorithms. The algorithms can reduce the energy consumption and extend the network lifetime.(3) We design more energy-efficient, lower algorithm-complexity k-vertex connected TC algorithms: GAFT/LAFT, which guarantee the network connectivity and fault-tolerance. Our algorithms solve the problem that the single-connected topology cannot support reliable network communication because of unreliable wireless and the dynamics in node sleep mechanisms and node faults. Furthermore, the algorithms meet two requirements at the same time: low algorithm-complexity and low energy consumption, which existing k-connected TC algorithms cannot satisfy at the same time. (4) Despite the advantage of k-connectivity algorithms, they avoidably result in the shortages, such as the limited fault-tolerance, high transmission power dense topology and so on. The thesis proposes the idea of topology recovery, and designs an algorithm LART to implement it. LART constructs a topology with light weight, low complexity and guarantees the network connectivity and fault-tolerant. Furthermore, LART constructs sparse topologies and achieves low energy-consumption and high network capacity.(5) TC, routing, MAC and physical layers unavoidably interact with each other. The cross-layer relationship has a great impact on network performance. We build an analytical model to study the cross-layer impact of TC.(6) It is still an important method to analyze the large-scale and randomly deployed network by simulation. Based on OMNeT++, the thesis designs a simulation framework R-Simulator, which costs lower simulation time and resource and is more realistic and closer to real networks than SensorSimulator. The numerical results in the thesis are all simulated and analyzed on the R-Simulator.The thesis introduces the close-loop method into TC in WSNs and brings new ideas into the area. The thesis designs it-connected algorithms to make the network fault-tolerant. Since this kind of algorithm has coherent shortages, the thesis proposes the topology recovery algorithm, LART. Finally, the thesis analyzes the cross-layer impact of TC, which brings TC into the holistic research framework of WSNs and provides a global perspective for TC research. The full thesis presents coherent and gradual research on TC in WSNs.