Fault Diagnosis in Body Sensor Networks

Abstract

Intermittent and transient faults are the largest source of failure for body sensor networks. In order to provide a method for detecting permanent, intermittent and transient faults, a distributed online solution is suggested. The diagnosis algorithm uses repeated testing in discrete time, and diagnostic messages are sent as the output of the routine tasks of the network. This work adopts a multi-tier telemedicine system where a star topology with the sensor nodes sending their data to Personal Digital Assistant (PDA) for data fusion is used. The system performs real-time analysis of sensor data and can generate warnings regarding an occurrence of fault, which allows fault isolation and system reconfiguration.