Dependable Network Protocols in Wireless Sensor Networks

This thesis is concerned with the dependability of Wireless Sensor Networks (WSNs). We propose an approach, inspired by the immune system, that allows individual nodes to detect, diagnose and recover from different failures by switching between different protocols using a multi-modal switching mechanism. A causal link between different failures in WSN is identified. Existing fault tolerance in WSNs approaches are examined. From the survey, it is identified that various attempts have been made to improve the fault tolerance of the communication protocol especially in the routing protocols. Although tests have been performed to evaluate the communication protocols prior to deployment, failures in WSNs are still being reported when deployed in real environments. A Systematic Protocol Evaluation Technique (SPET) is proposed and applied to evaluate the dependability of the proposed multi-modal protocol and reduce the uncertainties in the experiment and to demonstrate the confidence in the measurements taken from experiments.