“Sensor fusion is at the very pinnacle of environment detection,” explains Michael Maile, head of Sensor Fusion for Mercedes-Benz Research & Development North America, located in Silicon Valley. “That’s what combines all data together. The goal is to create an image of the environment that is as precise as such an image can possibly be.” This is a big challenge out in the heavy traffic of the city centre, as the vehicle must not only differentiate all the various obstacles on the streets, but also identify them. Children on the pavement, adult pedestrians, joggers, cyclists and motorbike riders, parked delivery vans and moving cars all have different and unique motion profiles.
The engineers and programmers at Mercedes-Benz are in the process of developing special algorithms to enable the vehicle to identify them and to assess and predict their behaviour. In the long term, the car’s perception skills must be just as good as those of any person driving on the road. Even the simplest processes that we find obvious as humans must be taught to a computer, and the exact way this must be done can be complicated. The city centre is thus a particularly complex challenge. Fast roads with more regulated, flowing traffic like motorways and highways are simpler, and much easier for the technology to manage.