ESI : to Showcase the Power and Efficiency of Parallel Computing at NVIDIA GTC Europe 2017

The ESI team will take this opportunity to share with users the capabilities of the combination of NVIDIA solutions with ESI's software Pro-SiVIC™, which has the capacity to simulate a real-time physically realistic 3D virtual environment and multi-technology perception sensors. The solution delivers key sensor models based on different technologies: camera, radar, laser scanner, lidar, ultra-sonic, as well as GPS, odometer and communication devices. It also provides a library of roads (urban road, highway, country road), traffic signs, and different lane markings. Parameter management allows quick and numerous changes, giving access to fully parametric scenes and scenarios to test full arrays of situations.

On the 2nd day of the event, Rodolphe Tchalekian, EMEA Pre-Sales Engineer at ESI Group will present a paper in the 'Deep Learning & AI: Autonomous vehicles' session. His talk on efficient training and testing of Advanced Driver Assistance Systems Artificial Intelligence (ADAS AI) with synthetic data will show the power and efficiency of ESI and NVIDIA technologies. When looking at machine learning, getting massive amounts of labeled and unlabeled training datasets becomes mandatory for preparing autonomous vehicles to analyze complex and unexpected driving scenarios. ESI uses NVIDIA's newest GPUs which bring the powerful and efficient parallel computing necessary to process data in real-time.

The team will demonstrate how Pro-SiVIC™ creates synthetic data to train and test any machine learning algorithm. Pro-SiVIC™ simulates multi-technology sensor systems data for numerous outdoor scenarios combining vehicles, obstacles, with pedestrians and weather conditions to cover a multitude of road situations and substitute countless hours of painstakingly recorded data. Using the NVIDIA DRIVE PX 2 platform, the ESI team can compare the performance of an already trained lane marking detection algorithm on a 3D Pro-SiVIC™ scene (simulated raw camera data) and a real video recorded from a car in similar conditions.

From now on, with ESI Pro-SiVIC™ and NVIDIA DRIVE PX 2, users can leverage the enormous power of Virtual Prototyping for the rapid scanning of multiple design solution, providing performance, robustness and reliability.

When? 10-12 October, 2017

Where? Munich, Germany

To register to the event, you can visit: www.gputechconf.eu

For more info, please visit: www.esi-group.com/company/events/2017/gpu-technology-conference

For more ESI news, visit: www.esi-group.com/press

For additional information, please contact: ­

Media Relations

Delphine Avomo Evouna

+33 1 41 73 58 46