Representing arbitrary sensor observations for target tracking in wireless sensor networks

Conventional object tracking in wireless sensor networks requires the use of predefined sensor types, which presents challenges when faced with adaptability and scalability constraints. This paper considers heterogeneous sensors and proposes a sensor-independent tracking framework. The Adaptive Grid Representation of belief Distribution (Adaptive GRID) provides a common data format to allow heterogeneous sensors to be treated homogeneously. The proposed framework has been evaluated by simulation, and the results demonstrate that the Adaptive GRID technique yields improved location estimation over a conventional occupancy grid approach. (C) 2016 Elsevier Ltd. All rights reserved.