An evidential fusion approach for activity recognition in ambient intelligence environments
Robotics and Autonomous Systems
| Hi-index | 0.01 |
Robotic middlewares increasingly allow the seamless integration of multiple heterogeneous robots into one distributed system. With the inclusion of ambient intelligence and pervasive computing, very simple devices such as WSN motes mounted sensors and actuators, rfid-tag attached everyday objects or smart objects should also be considered as part of this incorporation. Humans should also be a part of this NRS since one of the main objectives of this integrated robot systems is to serve humans. We argue that this inclusion of motes like tiny sensor-actuator devices, simple everyday objects and humans as part of the NRS brings many advantages when tasks are performed by cooperation. In order to achieve the goal, presenting all them in a robot middleware, my research (PhD) study has been started since Aug 2006. This paper represents our approach and achievement till today. In this study, models are inherited as well as proposed depending on the existing works. As the proof-of-concept, an illustrative experiment demonstrates how tiny devices and everyday objects become part of a NRS by adapting the proposed/inherited techniques, during this study, and cooperate to perform complex tasks.