Xilinx and Moshi Intelligent Launch Complete Hardware and Software Solutions to Promote Automotive Front-View Camera Innovation

New programs are underway in the $2.4 billion [1] forward-view camera market, focused on delivering comprehensive solutions that help automakers get to market faster

Xilinx, Inc. (NASDAQ: XLNX), the leader in adaptive computing, and Motovis, a provider of embedded AI for autonomous driving, today announced that they are collaborating on a solution for the automotive market that combines the Xilinx Automotive (XA) Zynq® System-on-Chip (SoC) platform with Motovis’ Convolutional Neural Network (CNN) IP specifically for vehicle perception and control of forward-looking camera systems. The solution builds on Xilinx’s corporate initiative to provide customers with a powerful platform to enhance and accelerate development.

Figure: Xilinx and Magic Vision jointly launch automotive market solutions to promote front-view camera innovation

Front-view camera systems are a key element of Advanced Driver Assistance Systems (ADAS) as they provide the advanced perception capabilities required for safety-critical functions, including Lane Keeping Assist (LKA), Automatic Emergency Braking (AEB) and Adaptive Cruise Control (ACC). The market-ready solution uses convolutional neural networks to achieve a cost-effective combination of low-latency image processing, flexibility and scalability to support a range of parameters required by the 2022 European New Car Assessment Program (NCAP) requirements.

“This collaboration is an important milestone for the forward-view camera market as it will enable automotive OEMs to innovate faster,” said Ian Riches, Vice President of the Global Automotive Practice (GAP) at Strategy Analytics. “The forward-view camera market represents a significant growth opportunity, with annual shipments expected to grow nearly 20% year-over-year between 2020 and 2025. The combination of Xilinx and MozVision will deliver a highly optimized hardware and software solution that will greatly meet the needs of automotive OEMs, especially as new standards emerge and demand continues to grow.”

The front camera solution leverages Magic Vision's CNN IP to scale across  28nm  and  16nm  automotive-grade Zynq SoC family devices. This unique combination combines optimized hardware and software functional divisions with a customizable CNN-specific engine based on Magic Vision's deep learning network, providing a cost-effective solution at different performance levels and price points. The solution supports image resolutions up to 8 megapixels. Now, for the first time, OEMs and Tier-1 suppliers can overlay their own unique algorithms on Magic Vision's perception algorithm set to bring differentiated and future-oriented designs.

"We are very excited to launch this new solution with Xilinx and bring our CNN front camera solution to the market," said Dr. Zhenghua Yu, CEO of Moshi Intelligence. "Customers designing systems with AEB and LKA require efficient neural network processing within the SoC with the flexibility to easily implement future features. With the unmatched efficiency and optimization provided by Moshi Intelligence's customizable deep learning network and the Xilinx Zynq platform's ability to support a dedicated CNN engine, we are meeting customer needs and helping to create future-proof designs."

Market forces continue to drive the adoption of forward-looking camera systems to comply with global government mandates and consumer watchdog groups, including the European Commission General Safety Regulation, the U.S. National Highway Traffic Safety Administration, and the European New Car Assessment Program. All three have issued formal mandates or strict guidelines requiring automakers to implement LKA and AEB features in new vehicles produced in 2020-2025 and beyond.

"With our comprehensive solution for the front camera market, we have expanded our automotive-grade solution portfolio to deliver cost-optimized, high-performance solutions to our customers. We are excited to bring this solution to market and drive the industry forward," said Willard Tu, senior director of Xilinx's automotive business. "Moshi Intelligence's expertise in embedded deep learning and its ability to optimize neural networks to address the challenging challenges of front camera perception provide both parties with unique advantages to gain market share and accelerate time to market for our OEM customers."

Xilinx and Magic Vision will  jointly speak at the Xilinx Adapt 2021  virtual event on September 15, 2021. Xilinx Adapt 2021 includes high-level keynote speeches, as well as partner and customer testimonials to help users unlock the value of adaptive computing. The Xilinx Adapt China series of events will be held online in November 2021. For details, please pay attention to the Xilinx Chinese website or Xilinx official WeChat.

About Xilinx

Xilinx is committed to supporting rapid innovation across a variety of technologies, from cloud to edge to endpoint, by developing highly flexible and adaptive processing platforms. Xilinx is the inventor of the FPGA and adaptive SoC (including our adaptive compute acceleration platform, or ACAP), designed to deliver the most dynamic computing technology in the industry. We work closely with multiple customers to create scalable, differentiated, intelligent solutions for the future connected world of adaptable, intelligent everything. For more information, please visit the Xilinx Chinese website: china.xilinx.com.

About Magic Vision Intelligence

Magic Vision Intelligence is an innovative technology company that integrates AI algorithm technology and advanced chips to provide autonomous driving products and solutions. Magic Vision Intelligence has successfully deployed a full stack of autonomous driving software and hardware systems based on precise environmental perception, sensor fusion, positioning, path planning and vehicle control algorithms, including parking and cruising, front view and surround view applications. Magic Vision Intelligence's products are widely used in mass-produced passenger cars and commercial vehicles by major OEMs and Tier-1 suppliers.

[1]  Strategy Analytics, ADAS Semiconductor Demand Forecast, 2025, Front View Camera Processor Market, August 2021