The cars are equipped with several sensors: long-range radar, medium-range radar, ultrasound, lidar, cameras. The various sensors provide data that must be merged into an overall image. This is called “sensor fusion” – a very important topic at conferences currently taking place, such as the Tech.AD Berlin, for example.
The difficulty for the engineers is to understand which sensor you can trust in a respective situation. A camera fails in fog or in backlight conditions while a radar sensor reaches its limits in other conditions. Robustness and redundancy are important terms that are heard time and again in this context. The challenge for the engineers is to create a reliable image of the vehicle surroundings from a multitude of potentially faulty sensor data.
The driving strategy, the so-called trajectory, must then be calculated from this image of the surroundings. Here, the car determines whether it is in the right lane, whether other vehicles are ahead or behind it, and how it must approach the next few meters ahead of it to avoid collisions.
In all these decisions, which contain a certain factor of uncertainty, the map serves as an important additional source of information. It tells the car, for example, exactly where the road markings are, so that the data it receives from the camera can be verified. This makes the maps indispensable for safety. Dynamic data, such as warnings of black ice, is also required to create the driving strategy in the car.
Interestingly, the prevailing opinion at Tech.AD was that the map data should be delivered “on demand”; in other words, from the cloud to the car. Permanently storing the data in the car and the updates associated with it would be a “nightmare”. This is entirely in line with the strategy we pursue with the Connected Electronic Horizon.
We would also like to share a few other impressions of Tech.AD with you: Laurie Cohen, a New Yorker who works for an investment fund, was impressed by the large number of highly qualified engineers and the level of the event. She is not aware of something similar to this in the US.
The development teams for autonomous driving are made up of experts from many different nationalities. The German automotive industry seems to attract engineers and software developers from around the world. They work together in mixed international teams in Germany. We spoke to a Chinese woman and an Egyptian who work for AVL Software and Functions GmbH in Regensburg. Together with colleagues from Australia, Portugal, Ukraine and India, they are developing solutions for autonomous driving. The German automotive industry seems to be a powerhouse that draws capable and skilled people from around the world like a vortex.