Block III: Industry Automation

In the last block, which began on December 9th , 2019 , the students were familiarized with the technologies of Industry Automation . The technical director is Shadi Saleh, M.Sc ..

The automotive industry is currently experiencing a major shift from conventional, driver-controlled vehicles to autonomous vehicles guided by artificial intelligence . The latter offers a safe, reliable, efficient and affordable solution to our previous mobility and thus redefines it in a drastic way. The techniques of " deep learning " have proven to be extremely successful in negating, in particular, perceptual restrictions for object classification, recognition and division.

However, it should not be overlooked that the " deep learning " method usually requires very specialized hardware , which must have a larger memory and the necessary computing requirements. The solution to this is either high-performance hardware or a less complex neural network , but a balance should be struck between the two for optimal results. So the right question is: What hardware and what type of network should you combine in order to achieve the maximum accuracy and speed of the deep learning algorithm?

The subject of Block III was to convey core aspects of the development of " deep learning " processes using embedded systems. Then there were various approaches to the development of " deep learning " processes and " computer-aided vision ", such as emergency braking assistance based on multiple object recognition and distance assessment, self-movement perception or vehicle recognition and location. Using the embedded GPU (e.g. NVIDIA Jetson Nano), the students were able to understand the development of the learning methods and gain practical experience.