Driver assistance systems (DAS) have now arrived on the scene. Even small and compact cars now feature complex, sensor-based systems such as collision avoidance systems. This is a development that will certainly intensify in the next few years. All systems currently offered already have one thing in common: They only offer assistance, while the control and responsibility remain with the driver.
It is now time take it to the next level. With the public’s attention, automobile manufacturers and suppliers are currently working on the vision of automated or autonomous driving:
- Mercedes-Benz tested an autonomous S Class car on an urban and rural route from Mannheim to Pforzheim.
- Audi has had its A6 automatically parallel park while the driver was standing next to the car.
- BMW has tested a driver-less 2 Series car in conditions when the vehicle is put to its limits or when drifting.
- Google has logged over nearly half a million miles on public roads with a fleet of autonomous vehicles.
All of these companies are working hard on the autonomous car. What challenges must be overcome so that today’s systems not only assist the driver but automatically drive the car without any action required by the person at the wheel? And what do customer think about this development?
The technical requirements
In order to regulate speed and distance to the next vehicle on highways with autonomous cruise control (ACC) while the driver still steers and controls the car, it is sufficient to have a manageable set of sensors and a rough map of the car’s surroundings. However, allowing a car to autonomously maneuver through highly complex and unpredictable traffic situations requires a precise and comprehensive model of the vehicle’s surroundings. Then there is still the imposing task of calculating the correct control impulses from this flood of data in real time.
The legal requirements
The Vienna Convention on Road Traffic of 1968 stipulates that the driver must always be in full control and is responsible for the behavior of a vehicle in traffic. Even if the car could drive itself better than a driver would: the technology currently is not allowed to assume this responsibility from the driver. A current topic of discussion is how this regulation could be adapted to include autonomous driving in the future. Connected to this debate is what kinds of legal liability implications would there be if such responsibility was removed from the driver altogether.
The levels of automation
Moving from assisted to autonomous driving will therefore not happen in one go. The German Federal Highway Research Institute distinguishes four levels which can be seen as a gradual development towards autonomous driving:
- Assistance: The responsibility still lies with the driver who must steer or brake the vehicle
- Partially automated: The responsibility still lies with the driver, yet the car may drive autonomously
- Highly automated: The responsibility in part lies with the vehicle and drives autonomously
- Fully automated: The full responsibility lies with the vehicle and drives autonomously
The difference is the degree of responsibility, and this is where the focus shifts to the driver.
The driver and her responsibility
To illustrate the differences in the levels of automation, the German Federal Highway Research Institute provides examples of systems: Highway assistant, highway driver, and highway pilot. The diffuse limits between these terms already presents a problem: As long as vehicle systems can properly control the car within their limits, the driver will not have to intervene and therefore not have to “fulfil his responsibility”. However, he may not have the ability to distinguish the levels of automation based on the mode of operation: They have the same feel.
Now it is the manufacturers’ turn to take up the challenge. They need to develop a concept for the displays and controls for such driver assistance systems which clearly conveys to the driver which mode of automation the car is currently in (“mode awareness”). On the other hand, he must be aware at all times of what kind of situation he is in ("situation awareness").
The car monitors the driver
However, not only must the driver understand her car, but also the car must also understand the driver. This requires a level of automation in which the driver has control yet the vehicle has to ensure that the driver does what she is supposed to: ACC adjusts only when the driver has fastened her seatbelt. A tracking system only steers the vehicle if the driver has at least one hand on the steering wheel. A driver on the highway would only drive automatically if it is certain that he could take the wheel at any time: This could be difficult if he’s reading a newspaper, sleeping, or has left the driver’s seat.
Therefore, autonomous cars in the future must be able to determine with much more accuracy what the driver is doing. The more closely the driver is monitored, however, the sooner he thinks he’s being patronized: A highway assistant is switched off if the driver moves his attention away from the traffic lane too often, as he would not have been doing what he was supposed to do. The fatigue wizard not only warns the driver but independently brings the vehicle to a halt on the hard shoulder, because the driver is too tired to continue.
What do customers think?
How do customers handle such powerful assistance systems? The modified roles of driver and vehicle lead to some questions, the answers to which will help decide whether the new technology is a success (on the market) or failure:
- What benefit might drivers have from various levels of automation?
- What concerns do customers have about the new assistance functions?
- How are the various systems used in practice?
- What limits to the system might the driver experience as a disturbance?
- How do drivers manage to distinguish the different levels of automation and to fulfil their responsibilities?
- How do drivers gain enough trust in the system to pass on the driving responsibilities to their cars?
- How much time is actually gained so that the driver can do other activities?
- What kinds of things does the driver want to do with this extra time?
- What kinds of changes to the driver’s seat will have to be made if he no longer is sitting in it to drive the car?
- How does the emotional relationship to the vehicle change if the driver becomes a passenger?
Spiegel Institut Mannheim has held an important role in exploring attitudes in consumers concerning driving assistance systems (DAS) and autonomous driving for several years. A meta-analysis of the studies already completed shows:
Generally positive: DAS increase safety and comfort.
- Warning: Systems that continuously intervene are still seen with skepticism.
- Comprehensive support: DAS are not only for individual, specific situations
- Maintaining control: It must be possible to overrule and switch off the DAS.
- Clearer status: The status of the system must always be clear as to be able to monitor the DAS.
- Intuitive operating concept: Even novices must be able to operate the DAS controls.
- Acoustic feedback can be annoying: Discrete warnings (signals, vibrations) are preferred.
- Autonomous driving: Concerns about losing the pleasure of driving or even driving skills
Spiegel Institut Mannheim offers a variety of research in the fields driver assistance systems and autonomous driving with its innovative research methods and tools:
- product-related long-term field trials of fleets of vehicles with automated functions through forums and online diaries using our tool logInsight®
- Focus groups
- Expert interviews on the subject of autonomous driving
- Representative surveys on the development of autonomous driving
- Lead user workshops, for example, on functional design and designing the display and control concept of automated features
- Innovators workshops to determine the requirements imposed on the functions of automated driving
- Development of new concepts and ideas in a concept lab
- Concept tests (driving and stationary clinics)
Spiegel Institut Mannheim is your experienced partner for designing intuitive display and control concepts for all levels of driver assistance:
- Holistic and innovative interaction concepts for concept cars for autonomous driving
- Front panel roadmaps for gradually phasing in different levels of automation
- Integration of new assistance systems into an existing display and operating concept
- Optimization of existing display and control concepts for all driver assistance systems
- Prototypical implementation of new display and control concepts
- Iterative tests of new display and control concepts with the relevant target group