Jungo announced an integrated next-generation CoDriver driver monitoring system (DMS) and occupant monitoring system (OMS) AI software using the Xilinx, Inc. 28nm Zynq-7000 SoC family, providing a comprehensive solution for the ADAS market.
This collaboration aims to provide a cost-effective, adaptable hardware and software solution. As convolutional neural network (CNN) algorithms continue to rapidly evolve, OEMs need an adaptable and mature platform to evolve, optimize, and enhance their CNN implementations onto. Jungo’s CoDriver software provides a sensor agnostic solution that tracks both drivers and passengers, including real-time distraction and drowsiness alerts, occupancy detection, children-left-behind detection, and much more – all from a flexible camera position in the rear-view mirror.
The Xilinx Zynq-7000 SoC contains a combination of programmable logic (PL), a deep learning processor unit (DPU), and CPU cores. By creating parallel processing pipelines within the PL, the Zynq-7000 SoC delivers the highest processing efficiency possible. With the DPU, a configurable computation engine within the PL, DMS and OMS deep learning models can become highly-optimized – offering up to 40x acceleration – thus providing OEMs a platform to efficiently process all requisite features for their DMS and OMS, all while balancing cost in a mature processing node.
The solution targets the growing demand for in-cabin sensing expected in 2024, as OEMs work to meet the New Car Assessment Program (NCAP) and European General Safety Regulation (GSR) standards targeting driver drowsiness and attention warnings.
“The tight implementation of CoDriver DMS and OMS software on the Xilinx Zynq platform offers a cost-effective solution with quick time to market,” said Ophir Herbst, Jungo Chairman. “The solution provides OEMs flexibility in optimizing their hardware resource use across multiple concurrent applications, driving costs down and increasing integration.”
“Expanding Jungo Co-driver software suite on the Zynq platform sets Jungo and Xilinx up for the explosive growth anticipated in the DMS and OMS markets,” said Willard Tu, senior director of automotive, Xilinx. “The adaptable Zynq-7000 family’s PL provides Jungo software upgradability to newly-developed AI techniques – therefore ensuring future competitiveness.”