A Framework for the Extension of DEVS with Sensor Fusion Capabilities

Abstract

Sensor fusion is a technique used to combine sensor data and improve the description of the physical property measured by these sensors. Sensors working alone could provide data that is erroneous, incomplete and uncertain. Several algorithms have been developed for sensor fusion, some of which are complex to understand and difficult to implement. The extensive list of algorithms often leaves designers confused about what choices to make as regards which algorithm is best and this discourages them from implementing sensor fusion. In this paper, we propose a generic sensor fusion framework to facilitate the implementation of sensor fusion algorithms using the Discrete-Event Modeling of Embedded Systems (DEMES) which is a model-based development methodology for cyber-physical systems based on discrete-event systems specifications (DEVS).