The so-called "AI car computer" is designed to steer automated cars through complex traffic situations that are new to the car. "We are teaching cars to navigate traffic independently," said Dr. Volkmar Denner, Chairman of the Bosch Board of Management, at the international industry meeting on the Internet of Things. Bosch sensors already enable cars to recognize their surroundings. Thanks to artificial intelligence, a car will also be able to interpret and predict how other road users will behave in the future. "Automated driving makes road traffic safer. Artificial intelligence is the key to this. The car is becoming smart," continued the Bosch CEO. Bosch also wants to work with the US technology company Nvidia to build the central vehicle computer. Nvidia is to supply Bosch with a chip on which the algorithms for vehicle movement generated using machine learning processes are stored. The AI car computer is set to go into series production at the beginning of the next decade at the latest.

The AI car computer knows what pedestrians or cyclists look like. In addition to this so-called object recognition, artificial intelligence should also make it easier for automated vehicles to recognize situations. Flashing cars, for example, are more likely to change lanes than non-flashing cars. This means that a self-driving car with AI can recognize and assess complex traffic situations such as a vehicle turning in front and take them into account for its own route. The computer stores the knowledge it learns while driving in artificial neural networks. Experts check the accuracy of the knowledge in the laboratory. After further tests on the road, the artificially generated knowledge structures can be transferred to any number of other AI car computers via an update.

The company only announced the establishment of a Center for Artificial Intelligence at the beginning of 2017. Bosch is investing around 300 million euros in expanding its expertise in this field.