This paper presents a systematic approach to the design and implementation of an energy-efficient multi-sensor network. The nodes of the sensor network form the basis of a sectioned Bayesian network that can be used to determine the state of the system being monitored. A key issue in the design of Bayesian networks for monitoring engineering systems is to ensure that reliable inference scheme about the health state of the system can be made by combining information acquired from each sensor in the system into a single Bayesian network. As the size of the network increases, aggregating information made by all the sensors becomes computationally intractable. Hence, sectioning of the Bayesian network based on functional or logical constraints allows computational efficiency in aggregating information and reduces overall communication requirements. Furthermore, an in-network data processing scheme, motivated by the concept of Dynamic Voltage Scheduling, has been investigated to minimize computation energy consumption through dynamically adjusting the voltage supply and clock frequency of the individual sensors. As a result, the processor idle time can be better utilized for prolonged computation latency, leading to significantly reduced energy cost and increased computational efficiency.
Efficiently utilizing the power available to increase service life of sensors is one of the key challenges in the design and operation of a wireless sensor network for system health monitoring. This paper addresses energy-efficient computation on the sensor node level by presenting an in-network data processing scheme. The scheme is motivated by the concept of Dynamic Voltage Scheduling (DVS), which minimizes energy consumption through dynamically adjusting the voltage supply and clock frequency of the individual sensors. Unlike the traditional approach where a uniform data processing speed is employed for all the sensors, the proposed scheme adjusts the speed of each sensor individually to utilize the processor idle time for prolonged computation latency. The advantage of such a scheme becomes increasingly evident when a large amount of raw data needs to be processed locally at each sensor to reduce the amount of overall data communication. An application model using vibration-base sensory nodes for machine health monitoring was constructed to test the new data processing scheme. Simulation has shown that energy saving of up to 29% could be achieved.