Was is autonomous driving?
The term “autonomous driving” is associated with vehicles that move independently without a driver or without being operated by remote control. This also includes self-driving transport systems and robots. This means that these vehicles are equipped with sensors, actuators and control systems that make it possible for the vehicle to move on its own. Alongside electrification, connectivity and car-sharing services, autonomous driving is one of the latest mega trends in the automotive industry.
Why autonomous driving?
The reasons for introducing autonomous vehicles are diverse. First and foremost, there is the need to increase road safety, mainly because the volume of traffic is increasing, and the demographic change is making it necessary to enable age-appropriate mobility (to reduce the risk of accidents among different age groups according to publications of the Technology Foundation Berlin). Then there is the fact that autonomous vehicles drive more efficiently and therefore reduce CO2 emissions. Finally, autonomous driving is of great importance to automation in the area of logistics, for example, in the form of self-driving logistics vehicles in production plants or in the case of “platooning” (see below).
What levels have been defined to achieve automated driving?
In order to clearly distinguish between the different levels of autonomy, the following levels have been defined:
- Level 0: Driver is fully active (steers, accelerates, brakes, etc.), there is no driving automation.
- Level 1: Vehicle offers driver assistance where the support systems such as cruise control, ABS, ESP intervene automatically.
- Level 2: Vehicle offers partial driving automation which takes over certain tasks without the driver having to intervene. This includes, for example, parking assistant, lane departure warning and congestion assistant.
- Level 3: Vehicle offers conditional driving automation whereby the driver does not permanently take control but after a certain warning time, is required to take control of the vehicle. The vehicle is able to take over the necessary functions (lane change, indicating, braking) itself.
- Level 4: Vehicle offers a high level of driving automation where the vehicle can permanently take over control. If the system is no longer able to master the tasks, the driver can take back control.
- Level 5: Vehicle offers full driving automation. No provisions are made for a driver.
When did autonomous driving begin?
Way before the term “autonomous driving” was coined, the automotive industry had been investigating the first technologies in this direction. As early as the late 1950s, Chrysler developed the “Cruise Control” application which keeps the vehicle’s speed constant regardless of the actual flow of traffic. This gave the impetus for further developments, culminating in driverless vehicles being tried out on a test site in the 1960s, the aim being to create reproducible conditions for tire tests. New developments in the 1990s such as radar or LiDAR made a significant contribution to making autonomous driving possible on the roads.
What special types of autonomous driving exist?
Commercial vehicle manufacturers also use autonomous driving systems under the term “platooning”. This is where several trucks are coupled together via an “electronic tow bar”. This is operated via the controls of the first truck which is steered by a driver, or semi-autonomously. Additional trucks follow the first vehicle synchronously and at a defined distance.
What does a vehicle need in order to drive autonomously?
In order to allow a vehicle to drive autonomously, a range of sensors (radar, LiDAR, camera, ultrasound) need to be connected to a powerful control device. These devices analyze the surrounding environment so that all objects, vehicles, people, etc. can be perceived. It is important to be able to detect and measure the direction in which they are moving along with their speed, size and distance etc. The sensors also have to detect traffic signals and traffic lights, etc. so that the current traffic regulations can be processed and complied with. In order to get a complete picture of the surrounding situation, all data from these sensors must be recorded as a whole.
The data obtained in this way is then processed by a complex algorithm to give an overall picture. Based on this, decisions can be made which require the control unit to have access to all necessary actuators (brakes, steering, acceleration) as well as all assistants in order to move the vehicle safely, even in critical driving situations. To obtain additional decisive information from outside the vehicle (e.g. control systems, Cloud or other vehicles), autonomous vehicles are equipped with what is known as “Car-to-X” technology. Information such as road surface conditions, collision warnings, pedestrians waiting at traffic lights, etc. is transmitted to the vehicle so that it can react to hazards at an early stage. In the event of system failures, the vehicle is equipped with redundancy systems which safely bring it to a stop in an emergency.
For example, in 2022, Mercedes was the world’s first car manufacturer to be granted approval to sell vehicles in Germany which were allowed to drive autonomously to level 3 at a maximum of 37 MPH (60 km/h). In this vehicle, 12 ultrasound sensors, 4 cameras, 1 stereo camera, 6 radar sensors, 1 LiDAR sensor and a highly sensitive GPS antenna are used. The sensor data collected is processed in a control device with huge processing power (comparable to approx. 15 Notebooks).
What impact does autonomous driving have on the vehicle’s design?
In order to operate a level-4 vehicle, the car only requires the driver’s attention intermittently and on demand. A level-5 autonomous vehicle no longer needs a driver, and only has passengers. The studies currently available show that this has a significant impact on the design of the vehicle and its interior.
For example, with a level-4 vehicle, the steering wheel can be folded down into the dashboard. Level-5 vehicles (“truly” autonomous vehicles) do not have a steering wheel at all. Various manufacturers have designed the vehicle’s interior with, for example, a U-shaped sofa or even a bed for long distance vehicles. In certain concepts, the driver’s seat can be converted into a mobile workplace, or all passengers can sit facing each other.
What about the issue of legal security in autonomous driving?
An important step towards legal security was the amendment to the 1968 “Vienna Convention on Road Traffic” recognized throughout the EU which established the basic concept that every vehicle moving in road traffic must have a driver. In 2016, this convention was amended such that, for the first time, vehicle systems were allowed to influence the way vehicles are driven (driver support systems, driver assistance systems or automated driving functions).
It is difficult to obtain full approval for autonomous driving throughout the world because the legal framework is so complex. An example of this is the fact that there is no driver and so the question arises as to who is liable in the event of an accident. Currently, in the case of autonomous driving, neither the owner nor the manufacturer can be held liable.
One key ethical question is the subject of intense debate. How does an autonomous (algorithmically-controlled) vehicle make decisions in the following situation: a pedestrian steps onto the road while a group of pedestrians is walking along the opposite side of the road, and a vehicle is driving towards them. Should the car swerve and drive onto the opposite side of the road or would it be acceptable for it to hit the single pedestrian in order to avoid a collision with the group of pedestrians?
What is the current state of development of autonomous driving?
An increasing number of reputable automotive manufacturers are offering middle and top class models with assistance systems up to level 3. Work on level-4 and level-5 vehicles is being intensively pursued. Different OEMs have – to some extent for several years now – already introduced suitable prototypes.
A few examples of reputable OEMs:
- Tesla has brought its “Autopilot” system onto the market. So-called Full Self-Driving (FSD) operates at level 2 and is used by vehicles to autonomously drive to destinations in the navigation system.
- Mercedes introduced an operational, autonomous luxury vehicle with an innovative interior design at the Mobility Show 2021. As already mentioned, the Mercedes S-class has been autonomously driving up to 37 MPH (60 km/h) since 2022.
- VW has developed an autonomous racing car (ID Pikes Peak) and intends to conquer the market for autonomous compact vehicles with its Sedric engineering design.
- Volvo is testing 100 automated XC90 vehicles on the Gothenburg beltway.
- In 2021, GM’s subsidiary tried out its “Cruise”, a level-5 driverless vehicle, on the streets of San Francisco.
What role does artificial intelligence play?
Systems currently being used already show that a significant part of the requirements are met if the environmental conditions for the vehicle are favorable. However, difficult conditions (such as snowfall, lack of road markings, unclear traffic situations) cannot be safely overcome with sufficient certainty with the current systems. This is where AI (artificial intelligence) will be used in the future. Some manufacturers use AI in connection with neuronal networks, and assess the data collected from test vehicles with the aid of “Deep Learning” algorithms. These measures increase confidence in making difficult decisions correctly (e.g. differentiating between people and objects).
How is the market for autonomous driving likely to develop?
If we are to believe the forecasts of well-known companies (Statista 06/02/2020), 10% of cars will be autonomous in 2030. It is expected, according to Bain & Company, that the percentage of autonomous vehicles will have increased to 40% by 2040. As a result, it can be concluded that this technology will be entering the mass market in approx 10 years’ time.