UPGRADE NEURAL NETWORKS FOR AUTONOMOUS DRIVING USING SIMULATION

Sensor are the eyes of an Autonomous Vehicles. They monitor 360° around it and capture continuously high-definition data of the surroundings. This data is only valuable, if it leads to an accurate analysis of the situation and subsequently, to the right driving-decisions. If we consider that computers are less powerful than humans at tasks that are not easily broken into simple steps like pattern recognition, the challenge to replace human by Artificial Intelligence could look significant.

At ESI Group, we strongly believe simulation could support you in upgrading Neural Networks (NN) for Autonomous Driving. As a substitute for hours of recorded data, Pro-SiVIC creates synthetic data to simulate the output from multiple sensor systems for outdoor scenarios that combine infrastructure elements, vehicles, pedestrians and other obstacles. In this presentation, we are highlighting how to leverage Pro-SiVIC’s advanced simulation capabilities to:

Reduce cost & lead time by enabling high-fidelity NN training data generation
Increase NN robustness in an efficient way by scaling up test coverage
Increase NN-based system development efficiency
Enable agile design-test loops and concept exploration to make NN safer before implementation

Upgrade Neural Networks for Autonomous Driving using simulation presented by Rodolphe Tchalekian

Reach out to a team of Future Mobility Experts

Please accept marketing-cookies to see this form.

In case you don't see the form or can't register please follow this link.