Dataflow-Oriented Provenance System for Multifusion Wireless Sensor Networks

Abstract

We present a dataflow-oriented provenance system for data fusion sensor networks. This model works best with net- works sensing dynamic objects and although our system is generic, we model it on a proximity binary sensor network. We introduce a network-level fault-tolerance mechanism by using the cognitive strength of provenance models. Our provenance model reduce the limitations of a sensor’s capability and decrease the error-prone nature of wireless sen- sor networks. In addition provenance data is used in order to efficiently build the dynamic data fusion scenario and to adjust the network such as turning of some sensors. In a fault-tolerant, self-adjusting sensor network, sensor data produce more accurate results and with the improvements, tasks such as target localization is more precisely done. One other aspect of our network is that by having computation nodes spread to the network, the computation is done in a distributed manner and as nodes make decisions based on the provenance and fusion data available, the network has a distributed intelligence.

Keywords: Multifusion, Wireless Sensor Networks, Open Provenance Model