Dependable Sensor Networks

Open Access
Wang, Guiling
Graduate Program:
Computer Science and Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
February 27, 2006
Committee Members:
  • Thomas F Laporta, Committee Chair
  • Piotr Berman, Committee Chair
  • Mary Jane Irwin, Committee Member
  • Srimat Chakradhar, Committee Member
  • George Kesidis, Committee Member
  • Runze Li, Committee Member
  • sensor networks
  • mobile sensor
  • sensor deployment
  • motion planing
  • sensor security
  • pairwise key establishment
Due to small size, low cost, and many other attractive features of sensor nodes, sensor networks have become adapted to a vast array of applications in both military and civil sectors, such as military surveillance, smart homes, and remote environment monitoring. To enable the usage of sensor networks in these applications, network dependability is the basic requirement. Specifically, a sensor network must successfully detect the phenomenon of interest, and transmit the generated data to the users reliably. However, this is a challenging task due to the harsh working environment and the extremely limited capability of sensor nodes. In this thesis, I have presented a number of protocols that are designed to improve the dependability of sensor networks. One fundamental requirement for a dependable sensor network is a suitable coverage, since a sensor network is deployed to sense a target environment, and coverage determines how well it can detect the phenomena of interest. Given such a requirement, the first step of my thesis research is to efficiently deploy sensor nodes to achieve required coverage. After the initial deployment, the dependability of the network may be degraded due to node failure, malfunction or compromise. To keep a sensor network dependable during its lifetime, I address both fault tolerance and security issues. In particular, the second step of my thesis research is two-fold: self-healing in response to node malfunction; and security mechanisms to protect communication. Extensive simulation has shown that our schemes are effective in improving the dependability of sensor networks in an efficient way.