Heavy stuff: serious computing muscle required for AV and ADAS simulation

Simulation is thirsty work for CPUs, particularly with a task as complex as autonomous driving. For developers, HPC has become a requirement. By Xavier Boucherat

Edge cases remain the autonomous driving (AD) sector’s Gordian knot. By definition, they represent the issues that developers can’t necessarily identify ahead of time. Nevertheless, one tool is helping to catch as many problems as possible: highly accurate simulation can place autonomous vehicles (AV) in environments which would be nigh-on impossible to replicate in real life, with complete control of weather conditions, traffic density, infrastructure quality and more, gifting developers the freedom they need to push AI and sensor technologies to the limit. Given that AD must deliver a significant safety benefit, nothing less than full flexibility will suffice….

Close
Close