Continental is developing Road Condition Observer, an algorithm designed to detect the typical features for the four road conditions of dry, wet, snow-covered and icy. The system uses sensors that are installed in the vehicle as standard and Cloud data to classify the road surface conditions.
“We use sensors available in the vehicle for the Road Condition Observer to gain information on the grip of the road surface,” stated Bernd Hartmann, head of the enhanced ADAS & tire interactions project group within Continental’s chassis & safety division. “This knowledge allows us to adjust the functions of advanced driver assistance systems to the actual road conditions. To prevent an impending collision, automatic emergency braking for example must be initiated considerably earlier on a wet road than on a dry one.”
In the future the Road Condition Observer will support the driver and will make the ADAS capable of detecting adverse conditions and to react suitably and in good time, according to Continental.
Hartmann believes that this capability will become more important in the future: “Automated driving also requires us to make judging the condition of the road technically possible for the system. An automated vehicle in particular must know if it is icy so that it can drive safely around the next corner.”
Traction action
In addition to the vehicle dynamics sensors ona vehicle, a mono camera is also used. The advantage of this is that the ESC is available in nearly every car as standard equipment and mono cameras are available in an increasing number of vehicles due to the growing distribution of driver assistance systems.
The Road Condition Observer classifies the road condition on the basis of the evaluation of camera images of the surroundings in front of the car and comparison with vehicle dynamics data from the ESC, knowledge of local and regional weather data (temperature, wiper activity and Cloud data), as well as tire behavior. In a subsequent step, a friction coefficient can be derived from this data.
“Where possible such a system should be implemented with the existing sensors. We have already been able to do this with wet road surface detection,” said Hartmann. “In addition to detecting wet roads, in the next stage there will be a vehicle dynamics model integrated in the ESC control unit that detects low friction coefficient values and can optimize the ESC actions accordingly.”
In the course of further development, Continental says the information from the Road Condition Observer will flow into a comprehensive 360° environment model that is prerequisite for a comprehensive understanding of the overall driving scenario. The environment model is generated through the fusion of different information sources: the integration of the road geometry, localization and traffic regulation recognition, model-based tracking of moving objects as well as the identification of free space play an important role.
