Cars learn to see, communicate, and think

In order for the assistance systems of the future to support the driver optimally in every situation, they need to know a lot ­ more than the driver himself. The component project Detection and Interpretation of the Driving Environment teaches these systems to see, communicate, and think.

To this end, specialists are optimizing and extending available monitoring systems to make them more exact, comprehensive, and reliable. Using laser, radar, and ultrasound sensors as well as video image processing, cars will acquire diverse kinds of data: on road conditions, weather, temperature; on traffic flow characteristics and the precise location of the vehicle. Faster than the human eye, they will detect the distance to preceding and following vehicles, obstacles in the road, or red lights at signalized intersections. At the same time, the systems must be carefully designed with a knowledge of the capabilities of driver and vehicle, in order to identify critical situations and decide whether, when, and how best to intervene. The enormous quantity of data accumulated by the vehicles must be assimilated and processed in a fraction of a second in order to produce the required information. The fusion of this data to conclusive information is crucial in allowing an assistance system to detect impending danger promptly and reliably and to intervene appropriately.

Faster and more effective interventions

Active safety systems are designed to help avoid impending accidents or at least reduce their severity. They can detect dangerous situations and warn the driver in time to avoid an accident, or in an emergency they can intervene and autonomously brake the vehicle to a complete stop. In the component project Anticipatory Active Safety, scientists are working on assistance systems to support a variety of lane changing and turning maneuvers.

For example, the intersection assistant will prevent drivers from accidentally running a red light or violating right of way rules and will protect pedestrians and cyclists. Thus, it is designed to detect everything occurring in front of, behind, beside the vehicle, and within the driver¹s blind spot; to recognize traffic signs and to support the driver by providing the appropriate information. If a situation is critical and the driver fails to respond appropriately or at all to a warning, the assistant intervenes to control steering, drive train, or braking systems while simultaneously stabilizing the vehicle. If a collision is unavoidable, vulnerable people such as pedestrians and cyclists are protected by adaptive safety elements such as flexible hoods or bumpers, which are also being developed in this component project.

Stop-and-go automatically

Accelerate, brake, speed up, slow down: Stop-and-go traffic can be tiresome and nerve-racking. In the future, Congestion Assistance will relieve the strain by taking much of the burden off the driver.

This development is based on known functions such as automatic cruise and headway control, which are being combined and extended so as to adapt the speed to the traffic flow and maintain a safe following distance dynamically. It supports the driver according to his requirements in acceleration, braking, and steering and automatically detects potential obstacles. In this way, the Congestion Assistance will not only avoid rear-end collisions, but will also improve the efficiency and smooth the flow in congested traffic. As a further positive effect, the more homogeneous flow will save fuel.

Really useful technology

Assistance systems must not only be intelligent, but also simple and intuitive to use, since their purpose is to decrease, not increase, the burden on the driver. Hence, scientists in a cross-sectional INVENT project are carefully investigating Driver Behavior and Human Machine Interaction.

This component project is systematically investigating how drivers react in critical situations to multiple stimuli and demands, and what perceptual and decision-making processes are at work. It is equally important to find out how drivers learn to operate new systems, what kinds of warnings are perceived and interpreted most rapidly, as well as what input could irritate the driver and even provoke errors. Evaluation of driving errors comprises the main criterion in a procedure for assessing traffic safety impacts of assistance systems developed in this project. The results of this research will serve as a basis for driver assistance systems to improve traffic safety designed to be as simple and self-explanatory as possible.

Clarifying legal issues right from the beginning

Just because an idea is technologically feasible does not guarantee that it will provide a benefit. In order to avoid wasting resources on lines of development doomed to failure, the cross-sectional project Traffic Impact, Legal Issues and Acceptance will evaluate the economic and business implications of the new technologies right from the start and investigate the needs of future users.

The methods applied to these problems include traffic simulation, customer questionnaires, market studies, cost-benefit analyses, compliance forecasts, workshops, and driving experiments. In particular, the experts are devoting considerable attention to conceivable legal conflicts. For example, who takes responsibility for cases in which an active safety system makes an autonomous decision and intervenes in a dangerous situation. Timely investigation of such issues is needed to allow rapid implementation of innovations in INVENT.