Presence Detection System for an elevator based on our 3D Ultrasonic Echolocation Sensor.
In this case, objects which enter an observed scene in front of an elevator are tracked in real-time to stop elevator doors from closing too early
Ultrasonic 3D Echolocation Sensors work by combining the time-of-flight principle of conventional ultrasonic sensors with triangulation and advanced signal processing algorithms. Therefore the sensor can locate objects in three-dimensional space and distinguish between multiple objects in a single scan.
The detection and tracking of moving objects within an area of interest could enable elevators to detect approaching people and therefore stop the elevator doors from closing too early while ignoring people not walking towards the elevator.
The system is capable of detecting objects within a predefined area, as well as tracking them to identify their trajectory. This enables the prediction of the path the object will take and enhances the reliability of the detection.
Through the clustering of our pointcloud data, new objects which entered an observed scene were tracked in real-time. The trajectory of these tracks allowed us to infer whether the objects were likely to enter a particular area of interest in the observed scene.
For this use case, our standard hardware and ROS drivers were used along with two new ROS nodes which performed the clustering and tracking of the objects. In the future this functionality could possibly be implemented within the sensor itself.
The use case could be vastly enhanced by using multiple sensors and through the further development of specialized algorithms. In addition, hardware modifications could be made to make the sensor more suited to the use case.