Barriers and challenges of AV deployment

LinkedIn +

From a developmental, infrastructural and regulatory perspective, the foundations for the rollout of autonomous driving have been laid, believes Sascha Semmler, head of innovation, autonomous mobility business area at Continental. Here he explains why.

Roughly 5-10 years ago, everyone looked forward to the introduction of fully automated and autonomous systems from 2020 through to 2025 – a very ambitious target. Then suddenly, 2020 approached and the introduction of these systems did not take place. The focus has massively shifted toward electromobility. As we look toward 2023 and beyond now, we can see a realistic path [for the introduction of autonomous vehicles]. Why? Because from an engineering perspective, we are able to tackle the critical use cases, which we were unable to do 5-10 years ago. We have a collaboration network with partners, and in terms of legislation, there are now standards and rules that allow us to drive on roads.

An important point, which should not be underestimated, is that these rules and regulations do not or should not fundamentally differ from country to country. This allows the development of one system that is in accordance with the legal regulations of many countries and can still be adapted to country-specific needs, instead of developing separate systems for different markets, which jeopardizes efficiency and standardization.

With regard to infrastructure, existing ADAS do not require any infrastructural changes at all. Furthermore, autonomous vehicles should be able to drive autonomously even without connection to the cloud, although there are additional benefits to doing so, such as knowing which potentially critical situations are approaching in the future, such as broken vehicles on the road, debris and intersections.

Physical versus virtual testing
There are two extremes related to testing – one is to test everything physically in the actual car or ECU, which is very expensive and slow; the other is to do intensive testing in the virtual environment. In reality, we combine these approaches in multiple ways. This allows much faster development of new modules and functions while ensuring strict safety measures.

Usefulness of artificial intelligence
When looking at AI, the amount of testing – given the increased number of functionalities and/or use cases – is increasing independently [whether AI is used in the development and the final product or not]. AI can help speed up the development cycles, but thorough testing must still be done. However, AI is a new key driver for enhanced functions and usability in the car and, therefore, fosters autonomous mobility.

From a developmental, infrastructural and regulatory perspective, the foundations for the rollout of autonomous driving have been laid, believes Sascha Semmler, head of innovation, autonomous mobility business area at Continental

Share this story:

Comments are closed.