Self-propelled vehicles have been the subject of debate on artificial intelligence in our everyday lives for some time now. However, one thing must not be forgotten when establishing these new technologies: road safety. The Berlin-based company Peregrine is striving to achieve this by using its technology to improve both traffic safety and operational processes. #ki_berlin met CEO Dr. Steffen Heinrich and talked to him about mobility services, intelligent cameras and their future and Berlin as a magnet for talent.
Mr. Heinrich, thank you very much for taking time for the interview! With the AI technology which you have developed road traffic should become safer. How did you get the idea for the product?
With the automotive industry, one of our most important industries is undergoing major change. The automation of vehicles and manufacturing, electromobility, connectivity and the development of new digital products and services are the major issues. The traditional automotive groups and suppliers are excellent in vehicle and engine construction, but have been facing massive problems in developing their own digital products and services for years. Software development in particular poses challenges for these players in digital transformation.
My co-founders Naja von Schmude, Jorit Schmelzle and I have spent the last ten years doing intensive research and development in the field of self-propelled vehicles, both in Germany and in Silicon Valley. We saw early on that camera sensors will become the central component of the perception system of driver assistance systems and self-driving cars. By 2025, every new car will have at least one fully networked camera. The video sequences from these cameras have a wide range of applications outside of driver assistance systems.
Visual information about the vehicle environment is very relevant for mobility services, insurance companies and the automotive industry itself in improving its own products and offers to customers. But also local authorities can make very good use of this data when optimizing infrastructure and in issues such as traffic safety and environmental protection.
We believe that a fresh start and double the pace are needed to ensure that the development of intelligent software and digital services also reaches the customer in Europe. It was clear to us that we could only build a new, innovative business model for the mobility sector as founders from outside.
How exactly does the technology work, and where in the transport sector is it actually used?
Imagine every car of a mobility service or the vehicles of a logistics provider had a smart camera behind the windscreen. A kind of virtual passenger with situation awareness to improve traffic safety, but also operational processes. Peregrine offers its customers exactly this "co-driver". The visual information significantly enhances telematics data so that processes can be precisely assessed and future developments can be better predicted.
It is essential that we already validate in the vehicle whether the image material contains information relevant for our services. Irrelevant image material is not stored or processed in the first place. This enables us to obtain only a small fraction of the data with a high density of information, and we save 99 per cent of the costs of collecting and storing the data. At the same time we apply the highest data protection standards to our data acquisition. This means that personal data in images such as faces and licence plates are alienated by an algorithm, making it impossible to draw conclusions about personal data.
What are the advantages of using your solution?
By using our software we actively prevent traffic accidents for example. In the event of an accident, we provide detailed information on the course of events. Our system makes it easier for fleet operators to constantly check the quality and improve it where necessary. For example, through individual training programmes for drivers, tailored to their routes and driving behaviour. In addition, for the first time our partners receive data-based decision-making aids on local traffic conditions. These include, for example, the risk along routes or driving profiles for so-called usage-based insurance products. In terms of technology, our data accelerate the development of driver assistance systems.
Your application is based on the use of data from traffic videos. What is the situation regarding data protection? Were you able to integrate this into your technologies without any problems?
From day one, conformity with the GDPR has been a central component of our architecture. We always try to collect only as little data as necessary for our services. When we collect information from image material, we use intelligent algorithms to alienate personal data. An anonymous data set remains, but with an accurate description of a traffic scene.
What was the decisive factor for you to settle with your company in Berlin? What advantages does the location offer?
In 2020, Berlin is still a magnet for international talent. This is a basic prerequisite for founding a start-up in the field of artificial intelligence. Out of 14 people, our team in Berlin has only two native Berliners.
But of course the question is justified, especially with regard to Naja’s and my work in Silicon Valley. Although the ecosystem there is certainly over-stimulated, it lives more than any other from an incredible network and knowledge exchange among investors and founders. With a view to Europe, however, Berlin also lacks successful tech start-ups in the software sector outside of SAP, Spotify or Wirecard. This deficit remains noticeable.
In recent years, the AI scene in Berlin has continued to grow. How do you feel about the status quo in terms of new technologies and innovations?
The AI scene is indeed growing rapidly. In 2012 there were only a handful of students with a focus on AI at the Berlin universities FU, HU and TU. Today this is definitely better.
But not every AI company necessarily creates its own IP. Often only the technology remains in use and still too often AI remains a sticker to enhance one’s own pitch deck. Socially, we are still very much at the beginning as far as the acceptance of AI systems is concerned. This will change when AI becomes more part of our daily life. AI systems, not only in medicine, biology and sports, but also in the financial world or the legal system are becoming more and more relevant, mostly as support for human colleagues. Successfully dealing with these topics requires that humans understand the algorithm to some extent and know its limits. The basis for this could be created by teaching more students in Germany basic knowledge about computer science.
How do you see Berlin as a location compared to other AI hotspots worldwide? Is it also a good place to operate internationally?
The basic conditions in Berlin and Germany are very good. University education is excellent. Industry and especially small and medium-sized enterprises are ideal sparring partners for AI companies: they have high professional expertise and data from real business. Many international competitors certainly envy us here. Compared to e-commerce solutions or software-as-a-service offerings, many AI applications are found in deep technical products. Here I see more experience and best practices internationally, especially in China, Israel and certainly also in Silicon Valley, than in Germany. This can be solved, but it requires very close and benevolent cooperation on both sides between companies based in Germany and AI start-ups. A disadvantage that AI start-ups share with other industries is the more difficult market access within Europe and the poorer scaling possibilities compared to the American market.
What are the possibilities for cooperation for your company in Berlin? Are there platforms or formats which have proven to be particularly helpful for you and which you can recommend?
Berlin offers great opportunities for founders to build a successful business. Starting with public programmes of the IBB or Berlin Partner, up to international accelerator programmes. We also have many successful founders in the city who are committed to ensuring that there are more success stories to come. Very good, partly self-organized meetups in the community help to exchange knowledge.
Thank you very much for the interview, Mr. Heinrich!