It is a nightmare scenario, but quite common: A vehicles rolls along through a residential area at low speed, but roadside obstacles - advertising columns, parking trailers, garbage containers and the like - block the view to the sides. Suddenly, a pedestrian appears behind such an obstacle, on collision course with the car. In this case, the assistant system has triggered the brake in time, the vehicle comes to a stop before it his the pedestrian.
The brake could be activated in this case because the pedestrian was equipped with a transponder, a combined of a radio receiver and transmitter which replies to certain specific signals - in this case to a pedestrian identification system in the approaching vehicle. To determine the exact location of the pedestrian, the system measures distance and direction of the person relatively to the car.
Within the Ko-TAG research project, TUM professor Erwin Biebl and his team devised a new approach to distance measurement which provides a result at an accuracy of a few centimetres and within microseconds. To trigger the measurement process, the car-based system transmits a unique code sequence. A transponder within reach modifies this bit pattern and returns it within a very precisely timing scheme.
To detect an approaching pedestrian, the driver needs to warned or the emergency brake has to be triggered even before the pedestrian steps onto the roadway. At the same time, the likelihood of an unnecessary full braking must be kept extremely low to allow car drivers to find the system reliable enough to use it. For this purpose, a very exact movement estimation is inevitable. Within their research, the TUM scientists were able to reduce the deviation to a few picoseconds. "Thus, we reach an accuracy of just a few centimetres in the distance measurement", said Biebl. "Along with the code-based method, this is the reason for the high performance".
A unique feature is that the system