NVIDIA unveiled Alpamayo R1 at this year's CES. This AI model is designed to control self-driving cars and supports Level 4 autonomous control (fully autonomous control within a defined area or conducive environment).
Alpamayo's unique feature is its vision language action (VLA) model, capable of making decisions based on inputs. Traditionally, autonomous driving systems separate object recognition and decision-making, limiting their ability to handle complex and rare scenarios. VLA allows cars to make decisions in less common and more complex situations.
Jensen Huang has previously expressed his vision for ChatGPT-based AI models to enter the physical world, and at this event, he indicated that self-driving cars would be among the first to benefit from this advancement.
At the event, Huang stated that the Mercedes-Benz CLA, launched this year, will use Alpamayo on the NVIDIA DRIVE DGX computer, but only at Level 2+.

No comments:
Post a Comment