Zach Anderson
June 11, 2025 12:14
NVIDIA wins the autonomous Grand Challenge at CVPR 2025 and shows the development of autonomous driving technology by the generalized trajectory scoring method.
NVIDIA has once again won the victory at the famous computer vision and pattern recognition (CVPR) conference, and has secured an autonomous Grand Challenge award for two consecutive years. This victory was presented in the specified intelligence of Horizon Workshop’s main event, the main event of the CVPR conference held in Nashville, Tennessee.
Development of autonomous driving technology
The focus of this year’s challenge, titled “Generalizable System,” is based on NAVSIM V2, a data -based simulation framework of autonomous vehicles (AVS). This competition aimed to push an envelope to develop smarter and safer AVs by encouraging research beyond the existing human driving data.
NVIDIA’s innovative approach, especially the Generalized TracJectory Scoring (GTRS), played a pivotal role in success. The GTRS method creates a wide range of driving trajectors that gradually filter to select the most optimal paths. This method uses a diffusion policy that is suitable for the environment to ensure compliance with safety and traffic rules using a combination of rough and fine trajectory.
New standard setting in safe driving
Participants who participated in the challenge had to create a driving trajectory in multiple sensor data in the semi -reaction simulation environment. The evaluation criteria included the measured safety, comfort and compliance using the expansion forecast driver model score. NVIDIA’s GTRS method is excellent in this evaluation and shows powerful adaptability in various driving scenarios.
This innovative system showed amazing generalization, achieved state -of -the -art results for challenging benchmarks, and is an important stage in autonomous driving research.
A wide range of influence of NVIDIA on CVPR
In addition to the autonomous Grand Challenge, NVIDIA was significant in CVPR 2025, and more than 60 papers were allowed in various areas including automobiles, medical and robotics. In particular, three automotive papers in NVIDIA were named Best Paper Award, and emphasized groundbreaking emphasis on stereo depth, understanding of monocular movement, and 3D reconstruction.
The meeting includes numerous workshops and tutorials led by NVIDIA experts, focusing on topics such as data -oriented autonomous driving simulation, non -language modeling and creation simulation. This session emphasized NVIDIA’s promise to develop autonomous driving and AI technology.
For more information, visit the NVIDIA blog.
Image Source: Shutter Stock