Visions of a future wherein our very environment is aware of our presence and can anticipate and support our unspoken desires are coming ever closer to reality. The ambient intelligence of such an environment is made possible through the combined functionality of numerous intelligent micro devices that are embedded within the environment itself. These devices sense data from the environment and use the information derived from this data to provide particular services.
The Project Motion Sensing develops an embedded system that contributes towards the vision of smart environments by providing services based upon intelligent object detection. This system provides a foundation for intelligent building control and makes possible functionality such as automatic summoning of an elevator or avoiding unnecessary elevator stops. Safety services are furthermore made possible, for example, to detect accidental falls in an elevator and to automatically alert management personnel. The embedded system’s clustering based object detection algorithm uses data observed by a time of flight sensor to pro-actively detect the presence and movements of individuals and groups of people. This context aware object detection algorithm has the small time and memory footprints required of a real-time embedded system and can robustly detect and track multiple objects despite noisy sensor data.