zapp2photo - stock.adobe.com

News

Snapdragon Summit 2024: Qualcomm gains mileage with automotive offer

Qualcomm accelerates automotive offering designed to drive success with leading OEMs such as GM, BMW, Li Auto and Mercedes as cars become new computing space

Joe O’Halloran
By
Published: 23 Oct 2024 10:45

On day two of its annual technical summit in Maui, Qualcomm revealed how it is adhering to the Hawaiian proverb “aʻohe hana nui ke aluʻia”, which can be translated as “no task is too big when done together by all”. It did this with the technology and OEM partner ecosystem it believes is necessary to drive its rapidly growing presence in the automotive industry, including taking video game technology into its digital chassis platform.

From an OEM perspective, Qualcomm showed the work it is doing with consumer car manufacturers GM, BMW, Li Auto and Mercedes, in addition to a marketing relationship with the Mercedes AMG Petronas Formula 1 racing team. While in terms of technology, Qualcomm announced an extension of its partnership with Google, Salesforce and Epic Games.

As he opened day two of the conference, Nakul Duggal, group general manager, automotive, industrial and cloud at Qualcomm Technologies, said that we are experiencing the transformation of cars which are more connected than ever, with high-speed connectivity in vehicles meaning driver and passenger experiences are more immersive.

Qualcomm, he said, was proud to have been part of this transformation journey over the past two decades, first connecting the car and then making the in-cabin experience richer, all the while designing for safety with a growing number of assisted in automated admin functions that continue to learn and improve through artificial intelligence (AI).

Duggal noted that the typical life cycle of a car is 10 years, meaning the importance of introducing to vehicles digital technology that can be refreshed and enhanced. In addition, unlike with mobile devices and laptops, the purpose of vehicles varies widely. A daily driver is different from a work truck driver, and is different again for those driving a convertible than an SUV. The connected car is also part of the technology ecosystem, meaning there is a wide variety of application and services needed to be relevant for these types of experiences.

“[Having] diverse users means architecting platforms for the modern vehicle is an exciting challenge. Ensuring that automakers have the latest consumer tech to design the experiences is critical, while ensuring that the product, technology, software and architecture is designed for safety, high reliability and the highest standards of quality,” said Duggal.

“The car is more than just mode of transportation. It is a becoming a seamless extension of our digital lives. It is a mobile living space, a space that is customisable on the fly… The industry is moving to advanced cockpit features, more intelligent and more immersive experiences that can be scaled across different types of vehicles. To meet the expectations of a wide variety of users, delivering premium audio-visual capabilities across multiple displays, requires automakers to integrate complex experiences, creating contextual relevance.”

Processing the massive amounts of data coming into the many information sensors now built into vehicles – fusing data from cameras, radars, lidars and location services – provides actionable information that can be used to inform a variety of applications, such as 3D mapping or premium navigation or safety alerts. Modelling a vehicle’s environment in real time is seen as a feat, requiring a significant amount of edge intelligence. AI can map point cloud data in real time to create an environmental model for a car.

For example, before executing a lane change, the car needs to know the distance to the car in front of it and how fast it’s going, and it must make similar predictions about every other vehicle on the road, their intent and predict their independent future trajectory.

“You almost have to be able to predict the future,” said Duggal. “The vehicle must then decide how to plan its course of action. To make these predictions and plan its route, your car [must have] a model of the world and [know what to] avoid, know where the safe spaces and overlay them in groups, learning decision-making and actuation.

“This is how the car operates – autonomous. The modern car is built on a digital experience and features are most advanced chips and technologies, products, software and services. This open and scalable, cloud, connected platform gives automakers technology building blocks to print the unique vision of the modern automobile life. It turns the car into a canvas that automakers use to extend their brand and their relationship with the driver.”

The core task, Duggal said, was to anticipate unique consumer needs and that the Snapdragon Digital Chassis allows automakers to create differentiated consumer experiences across every vehicle, bringing them to market and upgrading them as needed or as consumers demand, adding: “[It’s a] computing platform. You have your phone, your laptop, but now you also have the car. This is basically going to be like a smartphone on wheels. That is the future.”

In form, the latest version of the Digital Chassis comprises a number of Snapdragon units, in particular the Snapdragon Cockpit Elite and Snapdragon Ride Elite. Automakers have the option to use the Snapdragon Cockpit Elite for advanced digital experiences and Snapdragon Ride Elite for automated driving capabilities. Through system architecture, automakers will have an option to combine both digital cockpit and automated driving functionalities on the same SoC, a capability available on Snapdragon Digital Chassis solutions.

The new Snapdragon Cockpit Elite boasts an Oryon CPU optimised for automotive applications, designed for 3x CPP performance increase compared with its previous generation to offer what is described as “game-changing” in-vehicle experiences. The processor supports infotainment and an advanced driver assistance system (ADAS) on a single SoC. The dedicated Neural Processing Unit (NPU), designed for multimodal AI, is targeting a 12x performance boost over previous cockpit platforms, enabling real-time external environment and cabin data processing. This supports live decision-making, adaptive responses, and proactive assistance, ensuring personalised in-cabin experiences.

The Snapdragon Cockpit Elite and Snapdragon Ride Elite will be available for sampling in 2025. Technology collaborations are underway with leading car manufacturers, including Li Auto and Mercedes-Benz AG, who will feature Snapdragon Elite tier automotive platforms in their future commercialised vehicles.

Looking to the future

In terms of connectivity, the platform’s modems support 5G and the latest versions to be introduced in 2026 will bring 5 Gbps connectivity into a car. Duggal told Computer Weekly that the amount of data that cars are downloading and uploading is increasing massively. He said that the company always plans for introducing a new modem platform for the car every 30-36 months, but it is now starting to build the last generation of 5G platforms, which will likely be introduced for the 2028-2029 time frame. With the rapid availability of standalone 5G  networks, the ability to have a wide variety of flexibility in a connectivity architecture was very important.

But the real gains will be a product of partnership and collaboration. In its work with the tech giant as part of a multi-year strategic collaboration, Qualcomm will combine its Snapdragon Digital Chassis and Google’s in-vehicle technologies – such as Android Automotive OS and Google Cloud – to demonstrate co-innovation and knowledge sharing.

Duggal said that this will empower automakers to innovate and accelerate development of connected services, producing a new standardised reference framework platform for the development of generative AI (GenAI)-enabled digital cockpits and software-defined vehicles, with vehicle-to-cloud infrastructure based on foundational technologies from both companies.

At the heart of the delivery of this framework will be Google AI, seeing use to create GenAI-enabled in-car experiences such as intuitive voice assistants, immersive map experiences and real-time updates to anticipate driver needs, powered by Snapdragon heterogeneous edge AI system-on-chips (SoCs) and the Qualcomm AI Hub, designed to simplify deploying AI models for vision, audio and speech applications to the cockpit.

Pre-integrated Android Automotive Operating System (AAOS) software and services are designed to enable customisable and responsive voice user interfaces and real-time driver updates using Google Gen AI use edge optimisations through Snapdragon SoCs. The Google Cloud hosted automotive software development will be optimised for Snapdragon platforms to increase developer productivity while reducing time to market for AAOS platform and services. The Snapdragon Connected Services Platform running on Google Cloud will support an API-driven connected services model and an upgradeable services infrastructure for onboard and offboard vehicle connectivity.

“The automotive industry is on the cusp of major transformation enabled by advancements in generative AI, automated driving solutions, and the ability to update software features throughout the platform’s lifecycle,” said Patrick Brady, vice-president of engineering at Google. “We look forward to continue developing innovative solutions for software defined vehicles, and to showcase the power of co-innovation in propelling the future of transportation.”

Another IT giant that is rapidly expanding its offer in the automotive space is Salesforce. The company’s vice-president of industry solutions and strategy, manufacturing and automotive, Tony Kratovil, said that while the company was essentially a CRM leader, its foundations are based on the platform concept for managing all customer interactions.

As the firm has evolved the platform over the past 25 years to cover a wide gamut of industries, it launched its automotive cloud offer around a year ago in partnership with Qualcomm, bringing what Kratovil said were unique automotive requirements into the Salesforce platform.

“We have pre-built frameworks for managing viewer networks, pre-built frameworks for managing the customer consumer experience, and ultimately we have added vehicle [information] into the data model of Salesforce at its core,” he added.

“If you think of the solutions that are built on Salesforce by some of the leading automotive OEMs in the world, it’s all about customer engagement, making that consistent from the time that the vehicle is delivered to the customer all the way through the lifecycle of the vehicle. This includes services like subscriptions, and a complete 360 view of all the interactions that happen in having the vehicle.”

One issue identified by Kratovil is that the OEMs in the industry have a long history of technological development and an equally large number of legacy systems in use. This means that building a new world of the software defined vehicle is “very challenging”, both from an engineering and a deployment perspective.

Kratovil believes that the joint solution presents an opportunity to merge the worlds of cloud and devices together, unlocking all the data that lies in various connected services. Examples could include providing offers around insurance programmes, driven by the behaviour of the vehicle or other interactions.

“If you think about the opportunity here, it’s massive – not just for the benefits for a customer perspective, but also the seamless interactions that are possible now across the entire automotive value chain,” said Kratovil.

“Additionally, this platform approach offers the opportunity to reduce vehicle development to supplier management complexity, not just within the vehicle, but across the whole value chain, [including] dealers, dealer systems, dealer interactions and customer systems.

“This foundation provides the opportunity to unlock higher levels of customer engagement and personalisation. It also provides a much more nimble and effective mechanism for integrating with the dealer network. Dealers and OEMs can now be on the same page together, providing a unified experience to the customer. Hopefully, this helps manufacturers future-proof not just their vehicles, but their integration frameworks with customers and their dealer and distribution networks, and lays the foundation for services, some of [which] we haven’t even thought of yet.”

A less likely example of co-innovation comes in the form of Epic Games. Emphasising the aspect of collaboration, Bill Clifford, the company’s vice-president and general manager of Unreal Engine and creator marketplaces, revealed that Qualcomm and Epic have had a long history together, working through optimised games such as Fortnite on the Snapdragon mobile, and is now bringing the Unreal Engine 3D computer graphics game platform to automotive.

“Everyone knows Unreal Engine is the best engine for making AAA games, but a lot of people don’t know that the engine is used across a whole lot of other industries, basically anywhere where there’s real time to be design visualisation or simulation or any type of immersive consumer experience,” said Clifford.

“Automotive was the first industry to pull us in. We started in the 2010s when Oculus came out and a lot of the auto manufacturers were very excited to play in that space. They wanted to build apps for their design centres, for design review and marketing experiences for consumers. And they needed a real-time rendering engine to do what they needed to do. They’re creating digital twins that are leveraged in the design centre for designing vehicles.”

In automotive, engineering teams are using Unreal Engine to train AI models for autonomous driving using synthetic data and to simulate the driving experience. Sales and marketing teams are using for car configurator apps and digital experiences in the physical world, and, increasingly, the software sees use in digital cockpits to create a UX through human-machine interface (HMI).

Epic and Qualcomm have announced the deployment of Unreal Engine into the Snapdragon cockpit, with Snapdragon SoCs complementing the Unreal Engine features. This gives automakers exclusive access to pre-integrated and optimised Unreal Engine features.

“Qualcomm and Epic partner a lot to optimise games and the automotive space. They’ve always worked really, well with our engineering team to optimise the hardware and software staff to make sure Unreal Engine was delivering the best output in automotive,” said Clifford.

“The other thing that is really compelling about the integration is that Qualcomm does a really [good job] in taking all the windows on digital screens and compositing them into one universal, Unreal Engine-powered scene so an experience can be orchestrated throughout the car seamlessly, which is a great reason to build this solution.”

Read more about connected vehicles

Read more on Internet of Things (IoT)