The Nvidia Drive Hyperion computing architecture (Drive Orin), which is used by Pony.ai, a prominent provider of autonomous driving technology, to power its Autonomous Driving Controller (ADC), has been sampled by clients with plans to begin mass production of the ADC in Q4 2022.
Pony.ai has been doing road testing since January 2022 to optimise the programme for low latency and real-time performance. The centralised computation and AI engine for Pony.ai’s autonomous fleets will be the Nvidia Drive Orin system-on-chip (SoC).
With single or twin Nvidia Drive Orin configurations, Pony.ai’s newest ADC can produce 254 or 508 trillion operations per second (TOPS), respectively. Drive Orin, the highest performing automotive-grade SoC in the world, is built to handle several deep neural networks and apps at once while maintaining stringent safety criteria to provide intelligent driving that is both safe and cutting-edge.
Vehicles can intelligently and safely navigate congested city streets, motorways, and expressways thanks to the ADC and Pony.ai’s autonomous driving software and technology. This includes overtaking automobiles, automatically changing lanes to follow navigation, following traffic, and merging during traffic jams.
It can also drive on mixed-use roads, manage on-and off-ramp merges, automatically pass electronic toll collection gates, and securely handle U-turns, recognise traffic lights, and unprotected junctions. It can also steer clear of construction zones and accidents.
One of the first companies in the AV sector, Pony.ai, developed a product line with a variety of configurations using one or more Nvidia Drive Orin processors and GPUs using the Nvidia Ampere architecture. This product line speeds up Pony.ai’s concept of a reliable, mass-production platform for AVs and allows scalable deployment across self-driving vehicles and robotaxis.
Pony.ai’s co-founder and CEO, James Peng, states: “By working with Nvidia to co-develop a turnkey computing platform, Pony.ai has deepened its knowledge of the features and benefits of Nvidia Drive to quickly shorten our time to market.”
Pony’s ADC’s remarkable real-time speed makes it possible for a variety of applications and deployments.
“Nvidia Drive gives the computing capacity to enable the complex AI necessary for Pony.ai’s self-driving systems,” says Rishi Dhall, vice president of automotive at Nvidia.
Pony.ai is thus in a position to develop a driverless transportation future that is safer, more dependable, and capable of improving throughout the course of its robotaxi and robotruck fleets.
Pony.ai co-founder and CTO Tiancheng Lou states: “We’ve stepped into a quick technological iteration phase to fulfil client demand as our Nvidia-powered ADC is frequently tested and further integrated at Pony.ai.”
We are excited to shortly start mass producing our unique ADC and supply it to a variety of clients that use autonomous vehicles.
Some of the essential hardware for Pony.ai is created in-house. This includes PonyAlphaX, the company’s fifth-generation autonomous driving system, which uses the heterogeneous processing unit based on the Nvidia processor. As of May 2022, Pony.ai has amassed 13 million (8 million miles) of actual autonomous kilometres in the globe as of May 2022, including more than 200,000 driverless kilometres (124,000 miles).
With cutting-edge sensors, Nvidia Drive computation architecture, and aesthetic and design elements for L4 automotive-grade mass production fleets, Pony.ai’s sixth-generation autonomous driving system design made its debut in January 2022.
The Toyota Sienna Autono-Maas (S-AM), a seven-seat hybrid electric platform for autonomous mobility, will begin road tests in China this year and be implemented within Pony.ai’s public-facing robotaxi operations in the first half of 2023 as the first model outfitted with the company’s most recent platform.
Pony.ai and Nvidia started working together to build Drive Orin-based systems in May 2021, using Nvidia’s high-performance SoC to quicken the development of Pony.ai’s hardware.