2020 saw “significant organizational growth and progress” in the autonomous vehicle market, according to the Autonomous Vehicle Computing Consortium, a global group of automotive and technology industry leaders cooperating to help accelerate the delivery of automated and assisted driving compute solutions.
AVCC members have finalized two key foundation documents, new working groups have kicked off work, and critical topics such as cybersecurity are now being addressed. Additional work and liaisons are kicking off in 2021, and new members are invited to help drive these initiatives forward.
Armando Pereira, president of AVCC, says: “The need for AVCC’s guidance for autonomous vehicle computing has gained importance over the past year, and we are proud of our technical committees’ work to unite and build consensus around industry recommendations.
“We more than doubled our membership and are on the cusp of publishing two of the first cornerstone technical recommendations that will be the foundation for industry work for years to come.”
When it launched in 2019, AVCC featured eight leading OEMs, Tier 1 suppliers and semiconductor companies: Arm, Bosch, Continental, DENSO, General Motors, Nvidia, NXP Semiconductors, and Toyota.
In the space of a single year, the AVCC grew 150 percent, as original members were joined by an additional twelve, including:
- Sandia National Laboratories
- SEI ANTech Europe
These members bring a wealth of expertise in automotive compute elements and interaction, imaging signal processing, and cybersecurity.
Fourth quarter 2020 marked the launch of the newest working group, Cybersecurity, which commenced with a valuable membership poll of current cybersecurity practices, standards and needs analysis to drive prioritization of the work.
This quarter, AVCC will release a set of technical reference documents for AV Systems Architecture and Image Signal Processing.
The Micro-Benchmarks Working Group will release their first of three documents in development, focused on AV Condition Benchmarking Recommendations, in 2Q21.
These recommendations will facilitate easier cooperation, expedite RFPs, and provide the groundwork for deployment-at-scale.
The ongoing goals of AVCC are to:
- Develop a conceptual compute platform architecture, its hardware requirements and software APIs to define an automated and assisted driving-vehicle ecosystem that uses processors, accelerators and other application-specific engines;
- Offer technical recommendations/specifications guided by the objectives of portability, interoperability, scalability and a balance between performance and cost;
- Identify relevant system level requirements in the architecture, hardware or software APIs; and
- Work with standards bodies to identify, leverage and influence the development of industry standards.
It is the intent of the Consortium to focus this work to synchronize deliverables with the start of vehicle mass production cycles.
Driving the future of autonomous vehicles
The path to delivering autonomous vehicles is long and complex. Significant challenges to deployment include:
- ultra-high-performance computing requirements within the power, thermal and size constraints of a vehicle;
- security and safety concerns such as firmware protection, threat modeling, privacy, and regional regulatory and standards compliance;
- excessive costs and complexity of offering proprietary solutions; and
- public adoption.
The AVCC understands these technological complexities and obstacles need to be overcome, and is bringing the industry together to deliver recommendations that address these challenges.