Task-oriented distributed data fusion in autonomous wireless sensor networks

The evolution of sensor node capabilities makes distributed data fusion possible in autonomous wireless sensor networks (WSNs) for various purposes. We propose a framework of task-oriented distributed data fusion, and investigate the assignments of heterogeneous sensors on nodes in the network, so that system performance can adapt the dynamics of tasks and the topology of self-organised networks. This work provides an approach to improving the fusion performance based on partial information from WSNs. Such a task-oriented autonomous wireless sensor network can be a part of the infrastructure for cloud computing through the Internet. A hierarchy of linguistic decision trees is used to map the distributed information fusion. The performance evaluation is done from five aspects, quality of estimates, computing scalability, real-time performance, data flow, and energy consumption. Four classic decision-making problems in the UCI machine learning repository are used as the virtual measures from WSNs to demonstrate the merits of the proposed system compared with the central fusion models.