Qualcomm, best known as the provider of chips that power many of the world’s smartphones that don’t have an Apple logo, have been growing their presence in the automotive industry. Qualcomm recently highlighted its Digital Chassis strategy and some of the automotive partners that are leveraging the technology.
For the past two decades, Qualcomm has been supplying semiconductors to the automotive industry starting with its cellular technology for in-vehicle technology.
In recent years, in-vehicle connectivity has moved towards universal installation and the San-Diego based company has been expanding its reach into other parts of the vehicle to grow its business and leverage the technology it has developed in the mobile industry.
The Snapdragon 820A debuted several years ago for infotainment systems and is used on vehicles from multiple automakers including Jaguar Land Rover.
The Snapdragon Digital Chassis encompasses all of the Qualcomm’s automotive offerings including both hardware and software. It includes:
- Snapdragon Car-to-Cloud – a software and services platform to enable automakers to develop new revenue streams
- Snapdragon Auto Connectivity – a full suite of connectivity solutions including 4G LTE, 5G, cellular vehicle-to-everything (C-V2X), wifi, Bluetooth and satellite positioning to connect the vehicle to the surrounding world
- Snapdragon Cockpit – the platform that drives the in-vehicle user experience including infotainment, gauges and visualization
- Snapdragon Ride – Advanced driver assistance systems (ADAS) and automated driving systems (ADS) including the Arriver vision perception system
Qualcomm already has a strong position in the connectivity area and has some cockpit business but that and the ADAS/ADS business is set to grow significantly in the next few years.
On the digital cockpit side, Renault, Honda, Volvo and several Chinese automakers are all adopting the Snapdragon platform.
Beginning with the replacement for the XC90 and the new Polestar 3 later this year, Volvo and Polestar’s Android Automotive based infotainment will be running on Qualcomm processors.
Later this year, Honda will also launch a new vehicle in North America that uses the combination of Android Automotive and Qualcomm before going global in 2023. This will be replacing the current system that runs an older Nvidia Tegra processors.
A three-way battle between Mobileye, Nvidia and Qualcomm is shaping up in the ADAS and ADS sector. All three are offering a combination of both high-performance processors and software stacks including advanced perception and drive policy systems.
While Mobileye has dominated the market for camera-based ADAS for more than a decade, many automakers are now adopting solutions from Nvidia and Qualcomm with either those vendor’s software or in-house developed code.
Qualcomm is officially launching the Snapdragon Ride Vision System in order to get its piece of the ADAS/ADS pie. As various new car assessment programs (NCAP) around the world are starting to incorporate more stringent tests of ADAS capabilities, automakers are aiming to improve the performance of those systems.
All major players are focusing on improving the reliability of vision perception, since these systems all rely on at least machine vision. Qualcomm is using the Arriver Vision stack it initially developed in partnership with Veoneer and Qualcomm is currently in the process of acquiring the Arriver business unit from Veoneer.
The vision system is designed to support multiple cameras including higher resolution 8 MP cameras. The Qualcomm software development kit is designed to be modular so that automakers and tier 1 suppliers can use either Qualcomm’s full stack or integrate the vision stack with drive policy stacks developed elsewhere.
The Ride Vision system can also integrate a variety of other components including parking assist, driver monitoring and map crowdsourcing.
Even though automakers can use the Qualcomm’s vision perception stack with their own drive policy software, Qualcomm is also working with OEMs to develop a common drive policy. While the core of this drive policy is intended to be used across multiple applications, elements of it can also be customized for specific use cases.
The first production applications of Snapdragon Ride Vision System are expected in about 2024.
General Motors GM will be expanding the use of Qualcomm for cloud connected platforms later this year and it will also launch new models that are expected to be among the first that use the Snapdragon Ride ADAS processor combined with GM’s own software.
In 2023, the new Ultra Cruise system that enables hands-free driving into urban areas will also run on Snapdragon Ride with both the ADAS processor and the optional AI accelerator.
Qualcomm announced at its recent investor day that BMW would also be adopting the Snapdragon Ride platform for its vehicles at the start of 2025 and will utilise the full Ride Vision system. Also, Renault is planning to adopt the entire Qualcomm Snapdragon Digital Chassis platform for vehicles beginning in 2025.
Other automakers leveraging various parts of the Snapdragon digital chassis are Hyundai, India’s Mahindra & Mahindra and China’s Jetour, Nio, Xpeng, WM Motor, SAIC and JiDu.
Interestingly, while most automakers are shifting toward these higher-powered processors and more centralised compute solutions, not all are going exclusively with one chip vendor. For example Volvo is using the Nvidia Orin for the centralised compute for ADAS/ADS, it is using Qualcomm for the digital cockpit.
This is the same for Nio and Xpeng. In the wake of the current production disruptions caused by chip shortages, especially for older silicon technologies, the industry is trying to accelerate the transition to these more modern processors.
The ability to shift workloads from one to another if to maintain fail operational capabilities if one part of the system goes down, is one of the advantages of going to a more centralised compute architecture with fewer higher powered compute nodes.
While this is more complicated with a mixed compute environment it can be done if the system is designed with virtualization.
Qualcomm is comparatively late to providing the ADAS/ADS compute relative to Mobileye and Nvidia and most of the near term programs were sourced before Snapdragon ride was available. As the next batch of vehicle programs are sourced, it’s possible that more will follow the leads of BMW, Renault and Mercedes-Benz (which is going with Nvidia from 2024) and harmonize their compute platforms.