Greek
EnglishDeutschFrenchEspañolItalianDanishSerbianHungarianDutchRomânăČeštinaPolishHebrew (Israel)Greek
 

European Association for Accident Research and Analysis

Banner03

 

Main AuthorBerthold Färber
Co-AuthorsBrigitte Färber, Annika Johnsen, Michael Popp, Thomas Hummel
Type of Mediapdf-document
Publication Year2017
PublisherGesamtverband der Deutschen Versicherungswirtschaft e. V.
(All images from the publication)

(Only abstract translated)

Parking assistants differ in many ways, ranging from simple acoustical or optical / acoustical displays over camera driven systems (rear-view and top view) to semi- or fully automated systems. The evaluation of the potential of these systems to avoid damages or accidents must take into account aspects of system performance as well as ergonomic aspects and surround viewing conditions of cars. For the evaluation of parking assistants two methodological approaches are presented and tested.

As virtual technique a virtual shadow print method using the tool “BLENDER” is presented in detail. The eye point is defined 77,5 cm over the unloaded driver seat. As height of potential obstacles 70 cm were chosen (small child on a bobby car). Several cars from subcompact class to SUV and luxury class were analysed.

For the analysis three viewing areas were defined:

  • Front view: length of forward blind zone (from engine bonnet) + width of A-column > 6.5 cm.
  • Side view: blind zone, measured from body on each side + mean of columns > 6.5 cm.
  • Rear view: length of rear blind zone (measured from tail) + width of rear columns.

The comparison of the different cars gives a distinct and comprehensible order with respect to viewing areas for parking and taxiing manoeuvres.

The experimental comparison of different cars varies assistance level and viewing conditions in a systematic manner. On the operational level the sight conditions of a compact car are systematically degraded to a SUV. Further a SUV with poor sight conditions is still more degraded to an (existing) car with very poor surround sight. Systems with generic surround view, rear vision camera, top view camera as well as semi- and fully automatic systems are compared. Parking manoeuvres in the experiment were chosen according to damage frequency of comprehensive insurance and casualty insurance. The study was performed with 60 subjects on a test track.

Essential results:

Parking assistants reduce collisions if they provide as minimal system layout a generic surround view (with optical and acoustical displays). An absolutely necessary feature is rear end surveillance with automatic emergency braking to avoid damages and accidents, especially for reversing out of a parking space.

Cameras support taxiing and parking in small parking slots, but they do not avoid damages. Top view offers optical surround view, helps to detect obstacles and is easier to understand. The comparison of semiautomatic systems (only lateral control) and fully automatic systems (longitudinal and lateral control) show clear advantages of fully automatic systems. Only fully automatic systems showed no collisions when backing into a parking space. For cars with poor surround sight other than these systems are not sufficient.

Whenever the driver is part of the control loop (assisted or semiautomatic) not only the distance to an obstacle but also the driven speed should be part of the algorithm for warning or intervention. This algorithm provides a driver centred warning strategy taking into account the reaction time of humans and can thus avoid or at least reduce parking damages.