Left behind occupant recognition in parked cars based on acceleration and pressure information using k-Nearest-Neighbor classification

One of the major causes of lethal or serious injuries to children in non-traffic accidents with cars is founded on the unattended left behind of them in parked cars. Therefore, Delphi's safety division is interested in the development of a low cost left behind occupant recognition, so that since 2008 different approaches for a reliable detection system are evaluated. One of them is based on high sensitive analogue accelerometers that monitor vibrations occurring at the car chassis. The investigations show a recognizable signal produced by human beings seated in a parked car which provides enough information to determine the occupancy state of a car. The presented contribution describes the additional use of a second sensor (pressure signal) input to improve the classification reliability by fusing the information of both sensing elements. This is illustrated at the k-Nearest-Neighbor algorithm as preferred classifier.