Seeing Machines Snags Numerous Partnerships to Advance ADAS
Seeing Machines is partnering with Analog Devices, Ambarella, Autobrains, and Omnivision to build vehicle monitoring systems for drivers, occupants, and pedestrians.
More CES news coming your way today, this time in the autonomous vehicle and advanced driver-assistance system (ADAS) spaces. ADAS markets are competitive landscapes. And while competition may drive innovation in complex fields like ADAS, it may be advantageous for companies to join forces to develop new solutions.
Driver monitoring systems can detect when a driver becomes drowsy or loses focus on the road.
This week at CES, ADAS company Seeing Machines, which creates driver- and occupant-monitoring systems (DMS/OMS), has strengthened its position in the autonomous transportation market by announcing numerous industry partnerships. In this article, we’ll look at three of these new partnerships with Seeing Machines to understand the company's direction and the future of autonomous technology.
Analog Devices: Monitoring Driver Alertness
On Jan. 3, Seeing Machines and Analog Devices (ADI) announced that they will be working together to help develop DMS/OMS systems that are based on infrared cameras. In contrast with DMS/OMS systems with traditional cameras, systems that use infrared cameras can monitor in the dark or other low-light conditions.
Seeing Machines will leverage ADI’s portfolio of infrared and high-speed camera connectivity solutions. Specifically, Seeing Machines will use ADI’s MAX25614 infrared LED driver and its Gigabit Multimedia Serial Link (GMSL) serializer/deserializer (SerDes) hardware to help build the hardware necessary for an infrared-based OMS/DMS system. These systems will be paired with Seeing Machines' FOVIO chips for integrated OMS/DMS functionality. More technical information can be found in the FOVIO brochure.
How Seeing Machines' FOVIO Driver Monitoring System Evaluation Kit works. Image courtesy of Seeing Machines. (Click image to enlarge)
On the software side, Seeing Machines will then employ its proprietary AI solutions for tracking eye gaze, eyelids, head, and body poses. The software will interpret the optical signals from ADI’s hardware to detect warning signs of driver fatigue or distraction.
According to the ADI and Seeing Machines, the collaboration will meet European Commission General Safety Regulations (GSR) and European New Car Assessment Program (Euro NCAP) requirements. Seeing Machines interior sensing is on display this week at ADI's booth, LVCC West Hall booth 4725, as part of ADI's set of CES automotive demos.
Ambarella and Autobrains: Low-power ADAS
Following the ADI announcement, Seeing Machines also announced a joint collaboration with companies Ambarella and Autobrains.
Together, the three companies are developing three unique safety systems, all of which can be hosted on a single SoC. The solution, which includes a forward-facing, eight-megapixel camera and a five-megapixel, cabin-facing camera, is an open platform that combines technology from each of the three companies.
The collaboration will use Ambarella’s SoCs. Image courtesy of Ambarella
Seeing Machines is contributing its embedded Driver Monitoring Engine (e-DME) software stack, which is the company's core driver monitoring algorithm stack for supporting both accelerated and non-accelerated compute options. The collaboration will integrate e-DME with Autobrain’s signature and self-learning-based software stack on hardware developed by Ambarella. Specifically, Ambarella is offering its CV2FS/CV22FS SoCs said to provide industry-leading performance per watt in OMS/DMS applications.
Together, this collaboration aims to provide a low-power solution for ADAS, allowing for OMS/DMS systems that can come in a small footprint at a low cost and a low design complexity. With a smaller system producing less heat, the companies hope to unlock OMS/DMS systems that are more flexible and allow for new camera angles.
Omnivision: A Collab Targeting Size and Cost
Finally, Seeing Machines announced a collaboration with Omnivision, a leader in digital imaging hardware.
These two companies took to CES this year to announce the successful integration of Seeing Machine’s software on Omnivision’s hardware. The two companies teamed up to run Seeing Machine’s e-DME software stack on the Omnivision OAX4600 system-on-chip (SoC) platform.
Functional block diagram of the OAX4600. Image courtesy of Omnivision [Downloads as PDF]
The OAX4600 SoC is an automotive-focused video processor that was previously developed in collaboration with Seeing Machines as well. The OAX4600 features Seeing Machine’s proprietary Occula neural processing units (NPUs) alongside a tri-Arm Cortex A53 subsystem. With the OAX4600, the companies integrated an automotive RGB-IR camera to provide a proof of concept for a cost-effective, small-footprint automotive monitoring system.
What are your thoughts on the new developments in ADAS and other autonomous-vehicle technologies? Let us know in the comments below.