Big news! SuperStar Future releases a new generation of high-level autonomous driving vehicle computing platform

On May 25, the 8th International Intelligent Connected Vehicle Technology Annual Conference (hereinafter referred to as "CICV 2021") hosted by the China Society of Automotive Engineers and the National Intelligent Connected Vehicle Innovation Center was held at the Beijing Yichuang International Convention and Exhibition Center from May 25 to 27. Dr. Liang Shuang, co-founder and chief technology officer of Super Star Future, was invited to attend and give a speech, and released Super Star Future's new generation of high-level autonomous driving vehicle computing platform NOVA30P at the main forum.

Dr. Liang Shuang said that for high-level autonomous driving, the computing power of the vehicle computing platform is a "necessary but not sufficient" condition. As an autonomous driving domain controller, the design of the embedded vehicle computing platform must not ignore the limitations of hardware and power consumption. The technical difficulty that cannot be avoided in the implementation of high-level autonomous driving is to deal with very strict restrictions and requirements in terms of cost, system power consumption and reliability while expanding computing power. He mentioned that "the design of a high-level autonomous driving computing platform is essentially an optimization problem under restricted conditions, which should be solved through the coordinated optimization of software and hardware."

Based on this, Super Star Future has designed and produced the NOVA30P platform, a highly energy-efficient, highly reliable, low-power, and highly flexible heterogeneous in-vehicle intelligent core computing platform, and can provide customers with supporting software and hardware design and collaborative optimization technical services. The new generation of high-level autonomous driving NOVA30P uses a heterogeneous hardware solution with high-performance processing chips and functional safety ASIL-D level MCU. The solution supports the access and data processing of up to 20 4K cameras, 3 laser radars, and multiple millimeter-wave radars, and completes real-time multi-channel and multi-target perception processing. It provides convenient deployment support and fully guarantees the security and reliability of the computing platform from the hardware architecture and basic software configuration.

NOVA30P, a computing platform for high-level autonomous driving

The new generation of high-level autonomous driving NOVA30P adopts multi-core heterogeneous embedded modular design through innovative underlying technology and architecture, and performs hardware-friendly model lightweight acceleration through automatic model optimization tool chain to achieve orders of magnitude improvement in real-time performance. With high adaptability, high flexibility, high energy efficiency, low power consumption, safety and reliability, it focuses on improving the system's energy efficiency ratio, power consumption, and low latency overall performance to ensure flexibility, safety and reliability.

Computing platform for high-level autonomous driving: NOVA30P

Dr. Liang Shuang introduced that through the optimized tool chain for neural networks and the customized acceleration IP library for non-neural networks, NOVA30P will provide users with strong support for efficient processing of autonomous driving domain algorithms. He also mentioned that this in-vehicle computing platform with a maximum power consumption of no more than 90W will be mass-produced to customers in July this year.

As a national high-tech enterprise that provides high-performance autonomous driving in-vehicle embedded computing platforms and full-stack intelligent driving scenario solutions, Super Star Future was jointly incubated by the interdisciplinary innovation of the School of Vehicle and Transportation of Tsinghua University and the Department of Electronic Engineering of Tsinghua University in April 2019. It has established strategic partnerships with well-known domestic OEMs, robots, and first-line brands of autonomous driving, including Xilinx, Desay SV, Inhomogeneous, Yutong Bus, Shaanxi Automobile Holding, Pudu, Poseidon, Chery, WeRide, etc.

At present, through the key projects of the Ministry of Science and Technology's "13th Five-Year Plan" and the 2022 Beijing Winter Olympics unmanned driving project, Super Star Future has deeply participated in the research of many important national topics and projects, including "environmental perception and decision-making algorithm development", "real-time vehicle embedded communication and intelligent driving computing platform and exciting driving demonstration application", and shouldered the important task of exploring the implementation and industrialization of research results. In the future, through continuous independent technological innovation and cooperation with upstream and downstream partners in the industrial chain, Super Star Future will continue to help my country's automobile industry grow bigger and stronger and make intelligent driving simpler.