Analysis approaches for predicting performance of wireless sensor networks

Wireless sensor network (WSN) system designers must carefully design protocols and adjust parameters to balance various aspects of performance under stringent constraints such as energy and communication bandwidth. It is of vital importance to quantitatively analyze the effects of system and protocol parameters on various aspects of WSN performance, and to design protocols or select parameters based on the analysis to have predictable performance. This dissertation explores the possibility of predicting performance based on quantitative analysis and design with guarantees in three key aspects of WSN performance.; The first aspect is sensing coverage. Some applications need various degrees of coverage for monitoring of different areas. The work presented here proposes an energy efficient protocol that provides the flexibility to enforce various degrees of sensing coverage with proven guarantee. The second aspect is tracking performance for a duty-cycle based WSN system. The tradeoff between energy consumption and tracking performance is thoroughly analyzed in the dissertation. This provides a theoretical foundation for choosing system and protocol parameters and predicting tracking performance based on these parameters.; The third aspect is the communication capacity for periodic streams. The dissertation presents a theoretical analysis which answers whether the rates of a specified set of communication streams can be satisfied by a network with a certain bandwidth. This analysis also provides clear understanding of the effects of communication ranges and interference ranges on the achievable rates.; The overall result of the dissertation is significant progress towards an analysis capability for WSNs. In addition, the methodologies applied in the analytical approaches presented here serve as good examples for future extension to more aspects of WSN performance for similar analysis.