Basic Knowledge Basics of Autonomous Driving - Part 2
How can we continue to move safely in the face of growing populations, overcrowded cities, and congested road networks? One approach is artificial intelligence - not least because human beings are the cause of most road accidents. After the first part of the six-part series showed how autonomous driving is changing the automotive industry, we now take a closer look at technological progress.
Imagine you can simply call a self-propelled car using a smartphone app, be picked up at home, board the bus and be conveniently transported to your destination. However, the path to this dream is still rocky: the sensor mechanisms, software and connection technology required for a completely autonomous vehicle are highly complex, expensive and not even completely defined.
Even if some technical challenges still lie ahead, the different degrees of automation of vehicles have already been defined. In 2014, the US Association of Automotive Engineers SAE classified autonomous vehicle technology. The results were set out in the standard "Taxonomy and Definitions for Terms Related to On-Road Motor Vehicle Automated Driving Systems J0316". At the beginning of this year, the standard was updated to reflect the latest developments.
The borders are blurring
The first two levels of autonomous functionality have been on the market for some time. Stage 0 of the SAE classification is easy to explain: Here the driver controls his vehicle continuously and is in no way supported by assistance systems. He steers, accelerates and brakes himself, keeps a constant eye on the traffic and copes with dynamic traffic situations (e.g. lane changes or overtaking).
Stage 1 was accompanied by the introduction of limited driver assistance systems. Responsibility remains with the driver. Automatic Cruise Control (ACC) and Lane Keeping Assistant (LKAS) are useful aids, but the driver remains in control - he must keep an eye on the traffic at all times and react to changes in the traffic situation (e.g. speed limits or road conditions) and any hazards that occur.
Following stage 1, the manufacturer's definitions differ in some respects. Tesla, one of the most hyped vehicle developers, claims that its autopilot technology already enables autonomous driving today. The company cites numerous examples where drivers did not have to reach for the steering wheel or press the accelerator or brake pedal during a driving situation. However, engineers developing autonomous driving technologies with other manufacturers tend to classify Tesla's technology between levels 2 and 3.
In real Stage 2 systems, so-called semi-automated driving is made possible by the interaction of various individual systems (including some of the above, such as distance control cruise control or lane departure warning). These assist the driver in steering, accelerating and braking. However, the driver must always remain vigilant and observe the traffic. He must also be able to handle dynamic traffic situations independently and therefore always keep his hand on the wheel.
Only with stage 3, which goes hand in hand with conditional automation, do vehicles take over traffic monitoring, and the car occupants can lean back a little. At this stage, vehicle systems can access context data like a traffic jam assistant and control the car without the driver having to do anything. Such a highly automated car can, for example, perform certain driving tasks on the motorway without human intervention. However, the driver must be prepared to take the wheel again at any time if the system detects a problem and is overwhelmed by the traffic situation.
Autonomisation is progressing
It becomes really interesting from level 4: pedals and steering wheel are only needed in exceptional cases, and the driver can completely relinquish control of the vehicle. The car manages journeys on certain routes completely independently, monitors the traffic and recognizes its limits in good time enough to achieve a safe state in accordance with the rules (e.g. heading for a parking lot).
The ultimate goal of autonomous vehicle technology - level 5 - is a car that is capable of handling every traffic situation independently. The on-board systems involved take over steering, acceleration and braking processes, monitor traffic and master all dynamic driving tasks without human intervention, taking into account environmental and road conditions.
The final two stages of autonomous vehicle technology can only be implemented with the aid of extensive sensor equipment. Only in this way can the entire vehicle environment actually be recorded and evaluated. The change from level 4 to level 5 requires a new technological framework with ultra-fast wireless communication. This requires next-generation networks that transmit data from other vehicles, road infrastructure (smart traffic lights, camera technology, etc.) and high-resolution maps with extremely low latency.
The classification of the stages on the way to becoming an autonomous vehicle shows where we stand and what still awaits us in the future. The exact description of the forthcoming stages will certainly be revised and adapted to new software/hardware developments. At the same time, it is still relatively unclear at what point in the future true autonomy will be achieved. The concept of the autonomous vehicle should be perceived as fluid, growing and changing - just like the technology on which it is based.
This article was first published in German by next-mobility.news.