Autonomous driving: What is it & how does it work? | MOIA

Significance, operating principles & state of development of autonomous driving

What is autonomous driving?

Cars that drive themselves as if by magic – that's probably the image you have in mind when you hear the words “autonomous driving”. By and large, that's exactly what it's all about. However, there are no supernatural forces controlling the vehicles, but a great deal of sophisticated modern technology instead.

Autonomous driving is the target phase of an ongoing development process that increasingly equips vehicles with automation functions. Currently, as a driver, you are supported by your car within the scope of the technical possibilities when driving. In the future, however, autonomous driving should be characterised by the fact that the human being permanently abandons the task of driving. The vehicle moves independently during this time.

Nontheless, currently there still is a legal requirement that there must be a person behind the wheel at all times. There are already systems that allow a very high level of automation when driving (e.g. during a traffic jam on the motorway). However, as soon as the car lets you know, you must be able to actively intervene again.

The 5 levels of autonomous driving

The path to autonomous driving is a process that passes through various stages of automation. A series of levels defined in the SAE standard J3016 was established for describing this. This standard was developed by SAE International. The organisation designs concepts of this kind for the automotive industry on an international level.

Level 0 – no automation: There is no automation here yet. A vehicle only does something when a person actively intervenes. You actively take over the dynamic driving task.
Level 1 – assisted mode: This is already very common. In individual situations, the vehicle supports the person at the wheel. This includes the very common technology of cruise control, for example.
Level 2 – assistance systems with partial automation: Here, further systems are added to support the person driving. The difference is that the system now also takes over some of the transverse and longitudinal guidance. One example is the lane departure warning system, which many vehicles already have and which has been mandatory for new cars since July 2022. Up to level 2, the driver is still required to keep an active eye on the driving operations. You must be able to actively intervene at any time.
Level 3 – conditional automation; the vehicle can act autonomously in individual driving modes, but the person behind the wheel is still necessary as a backup. As soon as the vehicle gives an appropriate signal, you must be able to actively intervene again. These systems are already deployable in isolated cases or to a very limited extent.
Level 4 – highly automated: The car takes complete control of selected driving functions without actively asking a person to intervene. No-one has to be compulsorily behind the wheel, but the vehicle may then enter a safe state in certain situations and stop driving. Example: At level 4, the vehicle should be able to independently drive onto a motorway, travel some distance on it and then drive off again. Afterwards, the car would end the journey in a safe way unless you actively continue driving. The car thus moves fully automatically until it reaches certain technical limits (such as complex traffic situations within a city).
Level 5 – fully automated/autonomous: From this level onwards, you are a passenger in the vehicle at all times. The technology will then be so sophisticated that a car will be able to process even complex situations in traffic on its own.

How does autonomous driving work?

For vehicles to be able to drive autonomously, they need clever systems that take on different tasks. Similar to a driving human, they need eyes, ears and a brain to process information. Sensors and computers are used for this purpose:

Radar: On the basis of radio waves in the radio frequency range, speeds or distances etc. can be recorded.
Lidar: These are laser scanners that use light waves to identify objects up to 400 metres away. In contrast to the radar, a lidar perceives objects in the environment much more precisely.
Cameras: They provide direct visual impulses to the system, enabling it to recognise lane markings or traffic signs, for example. However, this does not provide distance information, which is why additional sensors are indispensable.

Autonomous driving works when these different sensor technologies are combined sensibly and their data processed accordingly. This requires computers with sufficient computing power – the brain of the vehicle. These make use of complex algorithms. They are designed to handle even complex traffic situations and always react appropriately based on the data received. This includes, for example, local differences in driving styles, so the algorithms may differ from country to country and would need to be adapted.

What are the advantages and disadvantages of autonomous driving?

There are several good reasons for autonomous driving. These advantages also essentially reflect the goals that the technology is intended to achieve in the future:

More safety in road traffic: Around 90 per cent of road accidents are due to human error. It is already apparent that assistance systems in vehicles can counteract this. Autonomous driving could prevent many more accidents in the future and, above all, reduce personal injuries.
Inclusion for people with disabilities: If vehicles drive themselves, a greater number of people can use them. This would give many people (e.g. older people or people with disabilities) further access to mobility and thus also to social participation.
Cost savings for public transport: Autonomous driving offers great opportunities for public transport in particular. At MOIA, for example, around 60-70 per cent of our running costs go towards driving staff. If vehicles are autonomous in the future, this would significantly reduce the fare for all passengers. It could also help to efficiently expand public transport in poorly served areas. With the help of autonomous vehicles, a mobility service can be expanded much more efficiently, even without additional drivers. Nevertheless, MOIA will continue to have mixed fleets for a long time to come, in which drivers are also urgently needed.
Climate and environmental protection: The use of electric drives can reduce the resource consumption of the transport sector. In addition, digitally connected vehicles can improve the flow of traffic – e.g. through automatically adapted braking and acceleration processes or intelligent traffic systems. The additional deployment scenarios for ridepooling lead to a reduction in motorized individual traffic and thus to lower CO₂ emissions.
Securing digital technology leadership: Through projects like ours in Hamburg, Germany has the opportunity to become the technology leader in the field of autonomous driving. Promoting this technology will enable the establishment of an entire value chain for autonomous driving in Germany, which goes far beyond ridepooling. Opening up a world market also promotes further innovations, secures thousands of jobs and strengthens the future viability of Germany as an automotive location.

Disadvantages currently arise primarily from the framework conditions to which the technologies are still subject. On the one hand, with a view to the stage of development: Although they are already very sophisticated overall, they cannot yet guarantee at all times that a vehicle will always do exactly what would be most sensible in the respective situation.

The legal requirements go hand in hand with this: With the law on autonomous driving (SAE level 4) passed in 2021, Germany is taking a global pioneering role. The framework created by the law offers the possibility of transferring innovative mobility technologies into regular operation and thus establishing autonomous driving on public roads. The law and the ordinance (Autonomous Vehicles Approval and Operation Ordinance - AFGBV) adopted in summer 2022, are therefore an important step towards strengthening Germany as a location for mobility and innovation.

The European Union will also adopt a full legal framework for type-approval of automated vehicles by the end of 2022. Initially, it will be limited to a small series of up to 1,500 vehicles in the entire EU (vehicle class M₁) or 250 vehicles per member state (vehicle classes M₁,M₂ and M₃). By the end of 2022, the quantity limit for vehicle classes M₂ and M₃ is expected to be adjusted to 1,500 vehicles. A regulation for large series, on the other hand, could follow from 2024, according to the EU Commission.

How far developed is autonomous driving? The example of MOIA

It’s not possible to predict with 100 percent reliability when autonomous driving in Germany and other countries will become an everyday reality. MOIA's goal is to develop an autonomous ridepooling system in Hamburg by 2025. This should then also be applicable to other locations. To this end, we work closely with our partner Volkswagen Commercial Vehicles. VWCV is providing the ID. Buzz AD, the right vehicle for the job.

Until 2025, intensive test drives will be implemented under real conditions with and without passengers - initially with safety drivers, in order to be able to optimally test the operational processes of an automated ridepooling service. In 2023, it is planned to test the ID.Buzz AD for the first time on the streets of Hamburg with VWCV. This will be followed by the first autonomous rides with a closed user group in 2024.

At MOIA, we are working to automate all processes so that the autonomous driving experience works best for you - from your request via the MOIA app, to boarding, to arriving at your destination.

Autonomous driving: Significance, functionality & development