The automotive industry is increasingly recognizing the business value of data from vehicles. To collect, analyze and monetize this data, manufacturers use the cloud. But the cloud also plays an increasingly important role in vehicle development. Marc Solsona Palomar from Amazon Web Services (AWS) promoted a holistic view of cloud and edge (i.e. vehicle) at the 26th Automobil-Elektronik Kongress in Ludwigsburg. Drawing on his experience as a developer in an end-to-end model - from cloud to edge device - Solsona Palomar evoked the central importance of interfaces in such a model. One benefit of such a holistic development approach is a shared comprehensive code base. AWS has now extended this model to its automotive arm, he said. He contrasted the cloud-based way of working with the approaches and processes typical of OEMs today for software development and integration for their vehicles' ECUs; these are largely opaque to OEMs, a black box. Together with factors such as variant diversity or regulatory requirements, this paradigm poses massive challenges for automakers; testing alone consumes a large portion of manufacturing costs. In addition, a large number of software developers are needed - right from the start of development - to ensure that the demanding deadlines set can be met. The software-defined car will magnify these issues, Solsona Palomar warned.
Save the date: 29th Automobil-Elektronik Kongress
The 29th International Automobil-Elektronik Kongress (AEK) will take place in Ludwigsburg on 24 and 25 June 2025. This networking congress has been the meeting place for top decision-makers in the electrical/electronics industry for many years and now brings together automotive executives and the relevant high-level managers from the technology industry to jointly enable the holistic customer experience required for the vehicles of the future. Despite this rapidly increasing internationalisation, the Automobil-Elektronik Kongress is still described by attendees as a kind of "automotive family reunion".
Secure your Conference Ticket(s) for the 29th Automobil-Elektronik Kongress (AEK) in 2024! Remember that the event has always been sold out for many years. Also, follow AEK's LinkedIn and check out #AEK_live.
To solve these challenges, Solsona Palomar proposed an approach he called "Environmental Parity." The promise of this approach: it would allow embedded systems developers in cars to benefit from the same agility and speed common to cloud developers. The way to get there: vehicle characteristics would have to be replicated in the cloud. In all of this, Environmental Parity relies on the fact that identical hardware architectures are used in the AWS cloud as in the vehicle, namely the Arm architecture. The connecting element between the two is the SOAFEE framework developed by Arm. It ensures that instruction sets and architectures are consistent in both worlds. However, this framework is not yet fully realized and available, Solsona Palomar admitted: "Our goal is binary parity between the two environments. Cloud-native in the car is a journey. We are not there yet."
Cloud also for development
Dipti Vachani from the Automotive Business Line of processor IP supplier Arm also referred to the SOAFEE framework in her presentation in Ludwigsburg. "Software-defined means it's in the cloud" she said. In her presentation, she laid out the benefits of a strategy she called "Develop in the Cloud, Deploy in the Vehicle." In addition to the benefits already listed above, she also cited the ease of deploying software on a wide range of different semiconductors. In addition, standardization on this chip range facilitates long-term support and offers advantages in terms of cyber security. Like Solsona Palomar, she emphasized the importance of environmental parity in this regard. "The software-defined (vehicle) is an element of the cloud," she said. "That means you have to have exactly the same environment in the cloud and in the end device, including functional security and real-time features."
Tool for connecting vehicle and cloud
Also presenting a tool that connects the vehicle and the cloud was Dirk Walliser, head of development at ZF. The basic principle of the Friedrichshafen-based company's new open software and data platform, called ZF Ocean, relies fully on the cloud. The development tool is based on Microsoft's Azure cloud platform and integrates ZF's development tools for embedded applications, simulation, Digital Twin and virtual validation - all "naturally" in the cloud. Using the system will not only speed up system integration, but also facilitate device and asset management, Walliser promised. Collaboration with customers also takes place via cloud-based processes; ZF's backend platform in the cloud can be interlinked with the OEMs' respective cloud or backend systems for this purpose, regardless of whether the customer also works with Azure or with other platforms, such as those from AWS or Google.
The digital twin in the cloud
A special Digital Vehicle Twin was propagated by Lars Reger, Chief Technology Officer of the semiconductor manufacturer NXP. The digital twin in the context discussed here resides in the cloud, and it differs fundamentally from the common digital twins that are usually used by the engineering departments of OEMs for IT processes as part of life cycle management. Reger promoted a Digital Twin whose life only begins when the vehicle is handed over to the customer. This Digital Twin would record data from the cars in accordance with data protection requirements, store it, discover conspicuous patterns and derive a benefit from it, which it would then return to the vehicles via over-the-air updates (OTA). This would allow the cars to learn in the background, so to speak - "just as people learn when they dream," Reger said. Examples of applications would include road quality mapping, performed in real time by the cars and relayed directly to the relevant administrative bodies. NXP sees its role in this interaction as a technology supplier for the communications computer as the central data hub in the vehicles and of specific algorithms, for example for battery management.