The person detector developed by professors Peter Nauth and Andreas Pech from FRA-UAS utilizes highly efficient and sensitive ultrasonic sensors, capable of discriminating a human being from an object in a situation where a collision is likely. The researchers utilised the same cost-effective ultrasound sensors that are already used for series vehicles in connection with parking assistance. The actual innovation lies in the pattern recognition algorithm applied to the signals of these sensors. In less than half a second, the algorithm generates a situational analysis that can be used to activate the respective protection systems.
Case the sensor detects that an impact with another vehicle is imminent, it activates systems like airbags and belt pretensioner. If the system however determines that the other party in such a collision situation is not a vehicle but instead a pedestrian, it initiates measures aiming at reducing the impact. Towards this end, several approaches are imaginable. “It would be feasible to reduce the rigidity of the vehicle’s body”, says Pech who focuses on data analysis and signal processing. “An alternative would be that the bonnet is heightened to mitigate the impact. Likewise, an exterior airbag could be provided which would deploy in such a case”.
In comparison with other pedestrian detection systems, the system at hand is relatively cost-effective, which would make it possible to equip even lower priced vehicles with such a system. The area of application for the system is not restricted to automobiles, however. It also could be used in fire fighting: In cases where smoke detectors trigger a fire alarm, the system from the Frankfurt researchers could be used to determine if human beings are still present in a burning house or flat.
For more information contact firstname.lastname@example.org