Abstract
Wireless sensor networks have been practically applied to many fields including industry, science and environment monitoring. A traditional wireless sensor network is likely composed of several distributed sensors and a limited number of sinks. The sinks gather the sensed data from sensors for further analysis and then make the actuators take corresponding actions according to the analyzed result. A sensor-sink based architecture is vulnerable to the sink failure problem. Meanwhile, if a sensor does not have a direct link with the sink, it must deliver sensed data by hopping through other sensors. The sinks may take a long time to collect data and to make the final decision. On the other hand, unexpected faulty sensors may result in an incorrect decision. Based on above, we propose an autonomous wireless sensor network to make a local decision based on the regional sensed data to raise the efficiency of action taking. Our proposed scheme uses a consensus based algorithm to improve the fault resilience. In this article, we prove the correctness of the proposed scheme and analyze the computation complexity. © 2010 ISNN.