Like virtually all other technology suppliers and OEMs, Bosch's environmental detection technology is based on several sensor principles - radar, ultrasound, and video technology. Because they come closest to the human eye in terms of structure and function, cameras play a central role in this - both in automatic driving and in driver assistance.
Bosch is now talking about how the company has succeeded in raising camera technology in cars to a new level of development: Thanks to the combination of a special multi-path approach with artificial intelligence (AI) for object recognition, the perception of the surroundings is much more reliable and road traffic safer, promises Bosch.
Surprisingly, the new Bosch MPC3 camera does not support the stereo principle, in which two camera eyes inherently create a spatial image. Instead, it is a mono camera whose built-in AI is intended to enable robust object recognition. The central component of the camera is the V3H-SoC from Renesas.
With its help, the Bosch engineers have taught the camera how to reliably detect whether a road edge can be driven on or not, even if there are no road markings, for example. The algorithms also improve existing driver assistance systems and expand their range of applications. In order to prevent collisions, it is conceivable, for example, to implement automatic emergency braking for various animal species. In addition, the reliability of triggering emergency braking increases because the camera can also detect partially concealed pedestrians. The new Bosch camera can reliably read text or numbers on traffic signs thanks to optical character recognition and display the information to the driver via a fade-in in the cockpit.
The MPC3 will be appear in production vehicles as early as 2019, Bosch said.