A multisensor data fusion algorithm using the hidden correlations in Multiapplication Wireless Sensor data streams

Abstract

While wireless sensor networks (WSNs) have been traditionally tasked with single applications, we have witnessed the emergence of multiapplication paradigms in the sensor network field such as Shared Sensor Networks and Virtual Sensor Networks. As the number of applications in a WSN increases, it also increases the WSN complexity and the amount of required transmitted messages. A major requirement in these networks is to save energy in order to extend their operational lifetime. Among the methods employed to extend network lifetime, Multisensor data fusion (MDF) is one of the most widely used. This technique can be very important when applied in a heterogeneous network (a common case for multiapplication WSNs). In a heterogeneous network data streams may come from very different contexts and may have distinct representations. Traditional MDFs are not able to identify these different contexts, since they are designed using an application-specific design for the network. In order to overcome this limitation, MDFs have to identify (hidden) correlations between sensors and to exploit such knowledge to monitor the behavior of sensors during their working life. Those correlations in a multiapplication environment could indicate that different applications may share similarities in terms of sensing that could be used in MDFs to achieve better results in terms of energy consumption and MDFs accuracy. In this paper, we propose a data fusion algorithm that identifies patterns in data streams generated by distinct applications in order to find the best correlations to apply MDFs. Our proposal is validated through simulations and tests on real nodes.