Performance measurement in wireless sensor networks using time-frequency analysis and neural networks

One of the major challenges in wireless sensor networks (WSNs) is reliability which is not necessarily a measurable single point; however, it's the result of a number of considerations including the sensor consensus, complexity of WSN systems, integration of different kinds of sensors and electronics, and the harsh environment. Evaluating reliability of WSNs is a crucial part for any kinds of applications in order to attain a better quality of service. Once the reliability of a WSN degrades the operators can take an appropriate action to recover it, such as replacement of sensors/electronics/battery, change of network configuration, etc. In this paper, we surveyed the literature which discusses the fault diagnosis of WSNs and presented a conceptual approach for performance measurement of sensors in terms of reliability.