DeepRoute.ai will integrate the NVIDIA DRIVE Hyperion AV computing architecture into its Level 4 autonomous driving solution. With this upgrade, DeepRoute-Driver 2.0 will meet automotive-grade requirements and advance the company’s plans to bring production-ready solutions to the consumer market. and advance the company’s plans to bring production-ready solutions to the consumer market. Road testing will start during Q2 of 2022, with mass production expected to begin in Q1 of 2023.

This integration will enable DeepRoute.ai to conform to automotive-grade standards more quickly by incorporating redundancy and diversity into its solution. DeepRoute.ai’s proprietary inference engine and algorithm superiority allows for energy-efficient computing performance – a major advantage for optimizing mileage range and workloads, and reducing vehicle design and validation costs.

Well-versed in NVIDIA’s software and hardware system architecture, DeepRoute.ai’s inference engine brings the DRIVE platform’s excellent performance into full play. Driver 2.0 will be compatible with different car models designed for various purposes, from ride-hailing and passenger vehicles to urban logistics services. 

DeepRoute.ai’s Robotaxis, SUVs, and medium-duty trucks will soon include NVIDIA DRIVE Orin SoCs

“We have gained significant traction since the launch of our affordable Driver 2.0 last December, and we will continue to foster our competitive advantage with NVIDIA DRIVE.” said Maxwell Zhou, CEO of DeepRoute.ai. “NVIDIA has played an integral part in the development of our technology since the establishment of DeepRoute.ai in 2019, and this advancement marks another significant step in the commercialization of our production-ready L4 solution,”

Driver 2.0 will be equipped with two DRIVE Orin SoCs that are each capable of achieving 254 trillion operations per second (TOPS). DRIVE Orin is engineered to fulfill safety standards such as ISO 26262 ASIL-D, as well as simultaneously process various applications and deep neural networks that are critical for autonomous vehicle decision-making. With the potential to need only one Orin chip in the future, the cost of the solution will further decrease.

DeepRoute.ai Driver 2.0 computing platform built on NVIDIA DRIVE Orin

“We share a mutual goal to achieve functionally safe and production-ready autonomous driving solutions. DeepRoute.ai has demonstrated L4 autonomy capabilities, handling complicated urban scenarios, and we look forward to supporting their ongoing innovation efforts as they move toward commercial deployment,” said Rishi Dhall, Vice President of Automotive, NVIDIA.

At CES 2020, DeepRoute.ai unveiled its computing platform based on the NVIDIA DRIVE Xavier SoC, offering excellent L4 autonomous driving at only half the cost of traditional solutions. Now leveraging the high-performance capabilities of DRIVE Orin, DeepRoute.ai can continue its R&D efforts--enabling safer, more efficient intelligent Robotaxi and medium-duty truck mobility solutions and operations.