Search

In this thesis work, a system for locating different occupants in a smart environment setting using a set of small, simple (binary) sensors and a robot (Turtlebot) is designed, implemented and tested. The sensors were chosen to be simple devices containing only "ON" and "OFF" status without any functionality to identify occupants. These sensors were ubiquitously installed in an office environment at Halmstad University. The Turtlebot is an assistant robot with a Kinect camera that supports the system to recognize the occupant. The system combines new inputs to previous information using a data association algorithm that makes predictions about the future location of the occupants. Preliminary results on a short time experiments of two different scenarios show that localizing two different occupants at the same time, using the proposed data association algorithm and face recognition can be achieved with more than 80% accuracy depending on the activities in the smart home.

In this thesis work, a system for locating different occupants in a smart environment setting using a set of small, simple (binary) sensors and a robot (Turtlebot) is designed, implemented and tested. The sensors were chosen to be simple devices containing only "ON" and "OFF" status without any functionality to identify occupants. These sensors were ubiquitously installed in an office environment at Halmstad University. The Turtlebot is an assistant robot with a Kinect camera that supports the system to recognize the occupant. The system combines new inputs to previous information using a data association algorithm that makes predictions about the future location of the occupants. Preliminary results on a short time experiments of two different scenarios show that localizing two different occupants at the same time, using the proposed data association algorithm and face recognition can be achieved with more than 80% accuracy depending on the activities in the smart home.