Download files
Abstract
Success of Wireless Sensor Networks (WSN) largely depends on whether the deployed network can provide desired area coverage with acceptable network lifetime. In hostile or harsh environments such as enemy territories in battlefields, fire or chemical spills, it is impossible to deploy the sensor nodes in a predeter- mined regular topology to guarantee adequate coverage. Random deployment is thus more practical and feasible for large target areas. On the other hand, random deployment of sensors is highly susceptible to the occurrence of coverage holes in the target area. A potential solution for enhancing the existing coverage achieved by random deployments involves the use of mobility capable sensors that would help fill the coverage holes. This thesis seeks to address the problem of determining the current coverage achieved by the non-deterministic deployment of static sensor nodes and subsequently enhancing the coverage using mobile sensors.
The main contributions of this dissertation are the design and evaluation of MAPC (Mobility Assisted Probabilistic Coverage), a distributed protocol for ensuring area coverage in hybrid wireless sensor networks. The primary contribution is a pragmatic approach to sensor coverage and maintenance that we hope would lower the technical barriers to its field deployment. Most of the assumptions made in the MAPC protocol are realistic and implementable in real-life applications e.g., practical boundary estimation, coverage calculations based on a realistic sensing model, and use of movement triggering thresholds based on real radio characteristics etc. The MAPC is a comprehensive three phase protocol. In the first phase, the static sensors calculate the area coverage using the Probabilistic Coverage Algorithm (PCA). This is a deviation from the idealistic assumption used in the binary detection model, wherein a sensor can sense accurately within a well defined (usually circular) region. Static sensors execute the PCA algorithm, in a distributed way, to identify any holes in the coverage. In the second phase, MAPC scheme moves the mobile nodes in an optimal manner
to fill these uncovered locations. For different types of initial deployments, the proposed movement algorithms consume only 30-40% of the energy consumed by the basic virtual force algorithm. In addition, this thesis addresses the problem of coverage loss due to damaged and energy depleted nodes. The problem has been formulated as an Integer Linear Program and implementable heuristics are developed that perform close to optimal solutions. By replacing in-operational nodes in phase three, MAPC scheme ensures the continuous operation of the WSN.
Experiments with real mote hardware were conducted to validate the boundary and coverage estimation part of the MAPC protocol. Extensive discrete event simulations (using NS2) were also performed for the complete MAPC protocol and the results demonstrate that MAPC can enhance and maintain the area coverage by efficiently moving mobile sensor nodes to strategic positions in the uncovered area.