The Snapdragon Ride Platform continues to push ADAS/AD forward.
Leading the way with open, scalable solutions and preferred collaboration models with OEMs and Tier-1s.
With the recently announced strategic cooperation with the Renault Group to co-develop high-performance computing platforms based on Snapdragon Digital Chassis Solutions for Renault Group’s next generation software-defined electric vehicles, Qualcomm Technologies has added another major OEM to its growing list of OEMs (e.g., Cariad, BMW, and GM) collaborating strategically to address very complex ADAS and automated driving applications. Collaboration is a critical factor for these complex applications and is necessary for successful implementation.
Automated Driving capabilities range from raw data from multiple sensors (e.g., multiple cameras, radar, lidar, V2X, and maps) processed at the edge or in a central architecture; then fused in a central compute application processor that runs an intelligent drive policy, motion planning, and high precision localization. Gone are the days when a trunk full of computers is an acceptable solution. These solutions need to be flexible, scalable, power efficient — and as active safety becomes a near-standard feature, be able to address multiple levels of functionality and price points across the entire vehicle lineup an OEM offers to consumers.
Snapdragon Ride is an end-to-end open, customizable, and scalable platform that enables automakers to collaborate. It includes a comprehensive family of SoCs ranging from open programmable solutions to Ride Vision products with integrated vision perception. These turnkey solutions include drive policy and crowdsourced maps to create driving features that address any scenario in highway and urban environments. In addition to the silicon solutions, Snapdragon Ride includes underlying tools and SDKs that enable OEMs and their partners to build ADAS and automated driving applications.
Just as important as the components of the Ride Platform is the collaborative working model employed to complement and enable OEMs and their needs to develop these applications. The Snapdragon Ride platform supports customers’ internal software development and offers pre-integrated and optimized solutions in the Snapdragon Ride Vision and Autonomous Driving stack solutions (refer to image 1 in Snapdragon Ride Platform Presentation).
Collaborative working models
As the value chain evolves into more collaborative models, choosing the right partnerships is imperative to align the core competencies of key stakeholders. As the trend continues to build on a collaborative design process, the roles and offerings of the value chain will also evolve. As automakers strive for differentiation and leverage of key assets, the needs of each stakeholder will also change, making the right partner choice crucial for success. With the key components needed to implement next-generation ADAS and automated driving systems, value chain members provide a wide range of expertise and offerings, and companies collaborate to create the entire system.
As OEMs begin to vertically integrate and bring critical functions such as driver policy, computer vision, and overall system design, this will drive how, and which parts will be provided by members of the collaboration. Which components OEMs contribute is not black and white; however, each has a unique offering and needs complementary components to enable them to execute their ADAS and automated driving strategy.
Snapdragon Ride SoCs
As OEMs strive for differentiation in ADAS and automated systems, and development of various components such as drive policy, fusion, and other components can be enabled through Qualcomm’s Snapdragon Ride SoC family (refer to image 3 in Snapdragon Ride Platform Presentation). Snapdragon Ride SoCs are scalable and software-compatible solutions that can address NCAP solutions to L3 driving and are paired with Snapdragon Ride SDK to optimize and enable in-house software development.
The Snapdragon ADAS SoC that is part of the Snapdragon Ride hardware platform is a purpose-built, programmable SoC from Qualcomm to address the needs of the ADAS market. It has hardware blocks that enable ADAS applications:
- Machine learning and vision processing blocks for AI based implementations
- Vision accelerators for post-processing applications
- Graphic accelerators for visualization and open GL support
- Video processor for streaming media support
- System safety manager to ensure ISO26262 compliance for applications
- Multi-core CPU for running general-purpose drive policy applications
Snapdragon Ride Hardware platforms
The reference Snapdragon Ride Hardware platform has extensive hardware and driver support for the most common sensors used in ADAS platforms, including cameras, lidar, and radars. To support a range of customers, it comes in multiple configurations (refer to image 4 in Snapdragon Ride Platform Presentation) supporting combination of single SoCs, multiple SoCs and accelerators.
All Platforms share a common architecture and several pluggable cards. The platforms are highly modular and can be adapted to several different use cases. Modules include:
- Camera SerDes
- Display SerDes
- PCIe Switch and peripherals
- Ethernet Switches, PHYs, and Peripherals
- Audio devices
Snapdragon Ride SDK
The Snapdragon Ride SDK is an enablement platform designed for auto OEMs and their developer partners to build safety-certified applications for ADAS systems on top of the Qualcomm Snapdragon ADAS SoC and the Snapdragon Ride hardware platform. It provides modules that provide abstractions to the underlying Snapdragon HW blocks and the driver software that maximize KPIs like latency, memory, and bandwidth usage, allowing application developers to focus on developing custom application logic. The application developer will be able to quickly develop and deploy their applications with modules that are pre-optimized for performance and latency. It is a great starting point for customers to use and prototype their applications.
Snapdragon Ride Vision System
Snapdragon Ride brings an industry-proven vision stack for the multiple sensors primarily focused on front and surround camera systems and a driving stack seamlessly integrated with an SDK and toolchain, comprised of multi-sensor fusion, map-based localization, planning, and prediction methods to support NCAP to L3 driving systems.
- Computer Vision (CV) software and Vision SoC co-designed by Qualcomm
- Supports L1 to L3 ADAS/AD
- 5th generation CV Stack
- CV Stack compatible with AD Map creation
- Hosts turnkey Qualcomm drive policy or OEM Drive Policy, 3rd party driver monitoring, parking:
- Automatic emergency braking
- Vulnerable road users
- Lane keep assist
- Intelligent speed assist
- Adaptive cruise control
- Speed limit information system
- Lane following
- Intelligent high beam control
- Rear emergency braking
Snapdragon Ride Autonomous Driving Stack
Snapdragon Ride Autonomous Driving Stack includes a multi-sensor environment model, low-level fusion, high-precision localization, and AI-supported prediction and planning modules to achieve a safe and robust ADAS and AD experience. Qualcomm collaborates with BMW as the lead partner to develop a complete driving feature portfolio from active safety to L3 automated driving. Through this collaboration, the Snapdragon Ride Platform offers OEMs a robust, validated driving stack in various geographical regions and includes a reference sensor configuration. Further, OEMs will benefit from regular updates through active learning and an extensive data management and simulation platform. With continuous improvement, the Autonomous Driving Stack will extend the Operational Design Domain (ODD) to manage most journey scenarios with the highest available automation level. Additionally, OEMs can use the Autonomous Driving Stack’s capabilities to boost their development ambitions of specific and differentiating driving features.