Supplier Continental uses AI-integrated cameras to detect human intent.
It is about the hazard detection and the informal communication in the traffic, what you at Continental has developed a new generation of cameras and their images with a Artificial Intelligence (AI) analyzed.
That’s another step in the direction Autonomous driving, because you want to teach the car at Continental, to assess the traffic situations correctly. The method by means of an AI is on the CES Asia presented, as well as the fifth generation of the multifunction camera from Continental.
The camera should 2020 be produced in series, with the AI algorithm to be integrated immediately. In this way, the system should be able to detect and predict the traffic situation. This is to set the vehicle at Continental on a par with the people and their perception, only that the car can then react faster. However, the prerequisite is that the AI recognizes both the travelers and the vehicle environment.
The prognosis should be recognized by the movements of humans, for example at the roadside. The AI then determines if the person is about to enter the street. This is done using the supervised learning method, which is called a driving school for the AI. The AI has to analyze a huge amount of data and it is constantly learning, even in use.
For training à la driving school, for example, radar and camera data from recorded rides are presented to the AI. This is how the system learns to recognize people and their intention. If the AI has it, it will check the “knowledge” of the AI. The more the AI has to recognize, the higher the computing capacity has to be.
One cooperates with Continental among others with Nvidia, Baidu and research facilities like the Oxford University, of the TU Darmstadt or the Indian Institute of Technology Madras in India.