A large-scale sensor network typically consists of numerous low-cost and resources constrained sensor nodes working in a self-organizing manner. Being embedded in the physical world, wireless sensor networks (WSNs) present a wide range of failures, due to environment conditions, hardware limitations and software uncertainties, and so on. Once deployed, the interactivity of a WSN greatly decreases, and network managers must investigate network behaviors with limited visibility into applications. Based on a real world environment monitoring sensor network system CitySee, this thesis addresses three key aspects for network parameters management in wireless sensor networks, i.e., injecting performance related time-varying metrics into each sensor node, and collecting these metrics...[ Read more ]

A large-scale sensor network typically consists of numerous low-cost and resources constrained sensor nodes working in a self-organizing manner. Being embedded in the physical world, wireless sensor networks (WSNs) present a wide range of failures, due to environment conditions, hardware limitations and software uncertainties, and so on. Once deployed, the interactivity of a WSN greatly decreases, and network managers must investigate network behaviors with limited visibility into applications. Based on a real world environment monitoring sensor network system CitySee, this thesis addresses three key aspects for network parameters management in wireless sensor networks, i.e., injecting performance related time-varying metrics into each sensor node, and collecting these metrics to enhance network visibility, providing more practical and efficient diagnosis models in wireless sensor networks. Through intensive simulations and real world implementations, I evaluate the performance of the proposed methods in a real system and verify the applicability.