ZF, Nvidia Team Up on AI System for Self-Driving Vehicles

ZF Friedrichshafen AG is working with chipmaker Nvidia Corp. to develop artificial intelligence (AI) technology for automated and autonomous driving systems for passenger cars, commercial trucks and industrial applications such as forklifts and other material handling devices.

The companies introduced the ZF ProAI processor, which uses Nvidia’s scalable Drive PX 2 AI computing platform, this week at CES (formerly the Consumer Electronics Show) in Las Vegas. ZF aims to launch series production next year and says the system can accommodate over-the-air software updates.

ProAI processes inputs from multiple cameras, ultrasonic sensors, radar and lidar to analyze a vehicle’s surroundings and help navigate around potential obstacles on the way to a pre-set destination. The system meets automotive-grade standards for operating in extreme temperatures, moisture and dust, according to the partners.

Nvidia’s Drive PX 2 system employs an advanced form of AI called deep learning and a Parker system-on-a-chip design that can process 24 trillion operations per second. Integrating data from the various inputs via a “sensor fusion” process allows the system to precisely identify its position on a high-definition map and monitor in real time 360° around the vehicle as it is driving.

ProAI also is configured for vehicle-to-vehicle/infrastructure communications. This will help create a “swarm intelligence” for vehicle fleets to improve safety and efficiency by sharing information and speeding the learning process, the partners note.

Nvidia is working with more than 80 carmakers and suppliers on self-driving vehicle systems. ZF, which offers and is developing a variety of driver assist and self-driving technologies, is said to be the first supplier to commercialize the Drive PX2 computing platform. ZF also plans to market the system to unspecified third parties.