Qualcomm mounts attack on ADAS compute platform market
Qualcomm’s new Snapdragon Ride aims to address the complexity of autonomous driving and ADAS by combining a high-performance hardware, artificial intelligence (AI) technologies and an autonomous driving software stack to deliver a comprehensive systems solution. It offers automakers a scalable solution designed to support three industry segments of autonomous systems, ranging from L1/L2 Active Safety ADAS for vehicles that include automatic emergency braking, traffic sign recognition and lane keeping assist functions to L2+ Convenience ADAS for vehicles featuring Automated Highway Driving, Self-Parking and Urban Driving in Stop-and-Go traffic and even to L4/L5 Fully Autonomous Driving for autonomous urban driving, robo-taxis.
The platform is based on the company’s Snapdragon family of automotive SoCs. The device contains several modular heterogenous high-performance multi-core CPUs, energy efficient AI and computer vision (CV) engines as well as a graphics processing unit (GPU). The company gives some technical data such as a performance of 30 Tera Operations Per Second (TOPS) for L1/L2 applications to over 700 TOPS; and all this at a power consumption of 130W for L4/L5 driving. While this does not seem very energy efficient at the first look, it does not require active cooling and enables the design of a self-contained, compact ADAS and auto-driving control computer. The Snapdragon Ride SoCs and accelerator are designed for functional safety ASIL-D systems.
While the company follows the big automotive OEMs in its assessment that the next wave of innovation will likely be in the L2+ Convenience ADAS segment, the hardware solutions utilized in Snapdragon Ride from a single system-on-chip (SoC) for an Active Safety ADAS system driven by regulatory mandates to a scalable architecture of multiple SoCs and dedicated autonomous driving accelerators allowing for fully autonomous self-driving systems.
The SoCs and accelerators at hand are built on the fundamental approach of heterogeneous compute capabilities designed for application requirements. These ADAS SoCs and accelerators effectively manage a large amount of data from onboard systems, leveraging Qualcomm’s next generation AI engines; image signal processors for camera sensors; enhanced digital signal processors (DSPs) for sensor signal processing; high performance CPUs for planning and decision making; GPU technology for high-end visualization and immersive user experience as well as dedicated safety and security subsystems across the SoC and autonomous driving accelerator.
Integrated as a part of Snapdragon Ride is a purpose-built autonomous driving software stack, a modular and scalable solution available to automotive OEM and tier-1 suppliers to accelerate their development and innovations. This software stack is offering optimized software and applications for complex use cases, such as self-navigating human-like highway driving, as well as choice of modular options like perception, localization, sensor fusion and behavior planning. This software infrastructure for Snapdragon Ride supports customer specific stack components to be co-hosted with the Snapdragon Ride Autonomous Stack components.
With the introduction of the Snapdragon Ride platform, Qualcomm clarifies its aspiration to secure a share of the growing market of automotive computing platforms for ADAS and Automated Driving. Hitherto, this market is dominated by traditional players such as Nvidia (who already has found wide acceptance among tier ones and OEMs), Xilinx with its ambitious and groundbreaking Zynq Ultrascale+ SoC family along with related software, Renesas with its R-Car platform (which is well entrenched not only among Asian carmakers) and, of course, NXP as the number one automotive semiconductor vendors with its increasing portfolio of AD and AI platforms.
Qualcomm is no novice in the demanding automotive business, but with the launch of the Snapdragon Ride platform it enters new terrain. This might pose a serious challenge to all the established players – all the more so as Qalcomm has enlisted the help of many experts in developing it. Among others, Qualcomm quotes high-level managers from Arm and Synopys for functional safety features and Infineon (with whom Qualcomm collaborated to ensure smooth interaction with Infineon’s Aurix safety microcontroller architecture. From the software camp, Blackberry QNX hinted that its hypervisor and operating system will run on Qualcomm’s new platform, and Elektrobit announced that it plans to integrate its EB corbos Autosar software with Snapdragon Ride.
Snapdragon Ride is expected to be available for pre-development to automakers and tier-1 suppliers in the first half of 2020. The vendor anticipates Snapdragon Ride-enabled vehicles to be in production in 2023.
Related articles:
Xilinx bolsters MPSoC family for more computing power
Renesas sets the course for autonomous driving
Automotive communications processor enables service-oriented architectures
ZF demos automotive supercomputer, autonomous minibus
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
