In the same week that the California Department of Motor Vehicles granted a permit to NVIDIA to start testing self-driving cars in the state, it was reported that Tesla, which uses NVIDIA’s chips to power the latest version of its Autopilot, signed a contract with Samsung Electronics to build an ASIC (Application-Specific Integrated Circuit) system for future self-driving applications.
While NVIDIA has been positioning itself as a supplier of the computing power behind self-driving systems, it wasn’t believed to have ambitions to build entire systems with sensors and controls. While it could still be the case, it would seem that Tesla and NVIDIA are both increasingly encroaching on each others’ area of expertise.
The chipmaker’s most advanced computer platform for autonomous and semi-autonomous features, Drive PX2, was first introduced in a commercially available vehicle earlier this year when Tesla announced that every car off its assembly going forward will be equipped with all the hardware necessary for self-driving.
The platform has proven popular with other companies developing self-driving technologies. It will power the self-driving capabilities of the vehicles in the ‘Roborace’ planned for events before Formula E races next season, and will also end up in some of Volvo’s test vehicles next year.
The company was testing its technology in its own test cars over the last year, but only on private property:
With the OK from the DMV, they will be able to expand their program. It became the 20th company to be able to test self-driving technology on California’s public roads:
- Volkswagen Group of America
- Mercedes Benz
- Delphi Automotive
- Tesla Motors
- GM Cruise LLC
- Zoox, Inc.
- Drive.ai, Inc.
- Faraday & Future Inc.
- Baidu USA LLC
- Wheego Electric Cars Inc.
- Valeo North America, Inc.
- NextEV USA, Inc.
- Telenav, Inc.
- NVIDIA Corporation
Most of those companies use variations of similar approaches to self-driving. Some are going directly for level 5 fully-autonomous capability, while others are trying to introduce more advanced systems incrementally leading to fully self-driving, like Tesla.
As for as the technology itself, Tesla and NVIDIA have similar approaches and both rely heavily on vision and machine learning technologies.