GM, Toyota join with Arm to define ‘real’ compute platform for autonomous vehicles
Silicon companies Arm, Bosch, NVIDIA, and NXP Semiconductors have joined with car makers General Motors and Toyota and automotive suppliers Continental and DENSO in a new alliance geared towards ‘making fully self-driving vehicles a reality’.
The new Autonomous Vehicle Computing Consortium (AVCC) was announced at Arm TechCon 2019 in San Jose as a collaborative effort for autonomous computing. The objective, ultimately, is to move prototype systems for autonomous vehicles to deployment at scale.
The group, which is open to new members, will seek, initially, to define a system architecture that reconciles the performance requirements of autonomous systems with the various physical, thermal, and safety limitations of vehicles. It will also develop the requirements for software APIs for “each building block” in an autonomous system.
New approaches to transport are emerging, noted Arm, spurred by innovation in autonomy, electrification, smart infrastructure, shared services, and connectivity. Advanced Driver Assistance Systems hint at the future, but the challenge to fit the kind of powerful compute systems that will enable automated vehicles into commercial vehicles is way off. The computing and automotive industries must collaborate to solve this challenge, said Arm.
Dipti Vachani, senior vice president and general manager of Arm’s automotive and IoT business, said: “Delivering autonomous vehicles is not something that one company can do on its own – it requires collaboration at an industry level. There are significant challenges to deployment that include ultra-high-performance computing within the power, thermal and size constraints of a vehicle to operate a large and sophisticated autonomous software stack.”
The other founding members of the group also made clear the of the task ahead, for the market to be able to deliver on the promise of autonomous driving and intelligent transport systems. “The path to delivering autonomous vehicles is long and complex,” commented Kamal Khouri, vice president and general manager of advanced driver assistance solutions at NXP.
Gary Hicok, senior vice president of automotive hardware and software systems at NVIDIA, said: “The hardware and software requirements for autonomous vehicles are enormous, requiring an energy-efficient, high-performance AI platform to process sensor data and achieve the highest levels of safety.
Michael Meier, director of engineering and product management for automated driving at Bosch, said: “As well as the development of hardware, there is a large and sophisticated autonomous vehicle software stack required.”