Mercedes-Benz and NVIDIA launch autonomous S-Class

Mercedes-Benz has unveiled a cutting-edge S-Class tailored for the AI era, constructed on NVIDIA’s DRIVE AV platform and engineered with a level 4 (L4)-ready autonomous architecture. This innovative vehicle merges the German automaker’s renowned safety emphasis with NVIDIA’s comprehensive autonomous driving technology.

This launch serves as a clear indication of the convergence of advanced AI, high-performance computing, and automotive-grade safety engineering in production vehicles. It underscores how collaborations between traditional OEMs and AI platform providers are shaping the future of automated mobility, spanning from luxury cars to robotaxis.

AI and safety integrated into the new S-Class

The new S-Class operates on Mercedes-Benz’s MB.OS and is outfitted with the NVIDIA DRIVE Hyperion architecture along with the full-stack NVIDIA DRIVE AV L4 software. This platform is crafted to facilitate future robotaxi and chauffeured mobility services while upholding a safety-first approach.

According to NVIDIA, the system employs end-to-end AI alongside a parallel classical driving stack that operates simultaneously. This dual-path strategy is a component of the NVIDIA Halos safety system, ensuring consistent, dependable performance even in intricate or unforeseen driving scenarios.

“Mercedes-Benz has been a trailblazer in the automotive industry, producing vehicles known for their exceptional craftsmanship and safety engineering,” remarked Jensen Huang, the founder and CEO of NVIDIA. “Five years ago, NVIDIA embarked on a collaboration with Mercedes-Benz to carry forward that legacy into the AI era.”

Beyond individual ownership, the S-Class is also positioned for shared mobility. As part of NVIDIA’s established partnership with Uber, the plan is to offer premium autonomous rides based on the new S-Class through Uber’s global mobility platform.

Navigating the complexities of real-world driving

NVIDIA DRIVE AV is engineered to transcend rule-based automation by analyzing intricate environments and deliberating through multiple options in real time. The software undergoes extensive training on NVIDIA DGX systems and validation through high-fidelity simulation utilizing NVIDIA Omniverse NuRec libraries and NVIDIA Cosmos world models.

At the hardware level, NVIDIA DRIVE Hyperion integrates redundant computing power, multimodal sensor diversity encompassing cameras, radar, and lidar, as well as software stack diversity. This design, known as “defense-in-depth,” aims to eradicate single points of failure, a crucial element for L4 autonomy.

The architecture primarily targets premium robotaxi and chauffeured services, where reliability, comfort, and trust are paramount. NVIDIA asserts that this approach enables the same platform to cater to both advanced level 2 systems and L4-ready automated driving in production vehicles.

By amalgamating extensive AI training, simulation, safety validation, and seamless integration with Mercedes-Benz vehicle platforms, the new S-Class extends the brand’s safety leadership into the era of autonomous driving. The outcome is a flagship model that exemplifies how AI-powered autonomy can be introduced without compromising on quality, comfort, or safety.