Wireless sensor networks have become increasingly pervasive with promises to fulfill many of our critical necessities today. One issue that has permeated sensor networks recently is mobility. Broadly, mobility in sensor networks can be categorized into two classes: Internal mobility and External mobility. Internal mobility is the class where sensors themselves can move from one location to another, while external mobility is the class where certain external agents (not sensors) move in the network. Both mobility classes have patent and significant impacts to sensor networks operation. However, being an emerging topic, a clear understanding of opportunities and challenges of sensor networks mobility is lacking today, and hence is an important need of the hour. In this dissertation, we make contributions in both classes of sensor networks mobility.
First, we study the issue of how sensors can use their mobility to enhance quality of network deployment. We define two representative mobility assisted sensor network deployment problems. In our mobility model, there are hard limitations in both sensors' mobility pattern and distance. Such limitations are natural due to constraints on sensors' form-factor and energy. We identify critical challenges arising in deployment under such hard mobility limitations. We then propose a suit of sensor mobility algorithms for our deployment problems, and demonstrate their performance using theoretical analysis and extensive simulations.
Second, we study the issue of external mobility in sensor networks from a security perspective. We identify a unique security threat in sensor networks called physical attacks. We define a representative model of physical attacks, wherein an external mobile agent (human being or robot) moves in the network detecting inherent physical/ electronic sensor signals to localize sensors, and then physically destroys them. We formally model such attacks in sensor networks, demonstrate their destruction potential, identify variations, and finally propose countermeasure guidelines against them.
With the emergence of mobility in wireless sensor networks, coupled with its significances, we hope that our work in this dissertation can provide strong foundations and further motivations for researchers to explore this topic that promises to revolutionarize sensor networks research in the near future and beyond.