An eye on Edge AI: Ambarella powers new Yandex driver monitoring camera; Hailo partners with Macnica
In another eventful week for artificial intelligence (AI)-powered edge solutions, Ambarella entered a new partnership with Yandex to integrate its next-generation video and image processing chip into the SignalQ2 LTE-enabled driver monitoring system (DMS) camera. Also, AI chipmaker Hailo has partnered with Japan-based semiconductor distributor Macnica.
Ambarella’s AI chip integrated into Yandex SignalQ2
The partnership will see the integration of the Ambarella CV25 edge AI vision system on chip (SoC) within the Yandex SignalQ2 camera.
The camera will then be installed in the Yandex.Taxi fleet of partner ride-hailing vehicles by 2022, as well as offered to other companies conducting similar businesses.
For context, Yandex.Taxi currently has partnerships with over 700,000 ride-hailing drivers in more than 1,000 cities across Eastern Europe, Russia, Kazakhstan, Norway, Israel, the Ivory Coast, Finland, and Ghana.
“The new Yandex DMS camera based on Ambarella’s CV25 will improve safety across the Yandex ride-hailing fleet, and those of other companies,” said Senya Pertsel, senior director of marketing for Ambarella.
The new Yandex SignalQ2 LTE-enabled driver monitoring system (DMS) camera is based on the Ambarella CV25 edge AI vision system on chip (SoC).
Source: Ambarella
“We are pleased to meet Yandex’s stringent requirements for high-performance AI processing, total security, and very low power, in their small form factor, windshield-mounted camera,” Pertsel added.
The novel Yandex SignalQ2 will now utilize Ambarella CV25’s real-time AI processing to confirm driver identity as well as state recognition (e.g. level of fatigue or attentiveness.)
In addition, the device’s Ambarella CVflow architecture is both low-power and high-performance, allowing the Yandex neural network-based DMS stack to be run efficiently, and its H.264/H.265 video processed at high speed (30 fps minimum).
“We pay great attention to safety in our service, which is why we developed the SignalQ2,” commented Nikolay Buldakov, SignalQ project leader at Yandex. “We chose the Ambarella CV25 SoC because it meets our performance and power criteria.”
The device can reportedly operate with video flows from up to three sensors, and the CVflow platform’s security functionality ensures the camera is running original firmware at all times.
Hailo enters distribution agreement with Macnica
The new collaboration will see Macnica distribute Hailo’s AI chips in Japan, including its Hailo-8 specialized AI processor for edge devices, and its M.2 and Mini PCIe high-performance AI acceleration modules.
The Hailo-8 features up to 26 Tera Operations Per Second (TOPS) and is built with an innovative architecture enabling edge devices to run deep learning applications that could previously only run on the cloud.
The move is aimed at enabling Hailo to expand its presence in the country, and particularly in fields like smart retail, smart cities, smart homes, industry 4.0.
“Working with Macnica will significantly strengthen our presence in the Asia-Pacific region, where interest in Edge AI solutions continues to rise,” explained Hailo CEO Orr Danon.
According to the executive, the new distribution agreement comes at a time of growing demand from Japanese customers of devices with AI capabilities at the edge.
“We were impressed by Macnica’s…strong relationship with customers, and we are confident that with our collaboration, we will further expand our offerings in the region,” Danon concluded.
Article Topics
Ambarella | chips | computer vision | edge AI | Macnica Hailo | processors | Yandex
Comments