“For the BMW Group, the use of industrial data is a key factor in translating artificial intelligence into value creation,” said Dr Franz Decker, CIO and senior vice president of the BMW Group. “By combining our engineering datasets with Mistral AI’s model training capabilities, we are building specialized AI which supports complex development tasks.”

Complexity and data volume in crash simulation

Each week, BMW runs thousands of virtual crash simulations, generating vast amounts of engineering data. Over time, this has resulted in a historical dataset of over one petabyte of crash simulation data that provides highly detailed insights into vehicle structures and material behavior, forming a unique foundation for training an industrial AI model.

“As Industrial AI becomes the new frontier for AI, we are proud to partner with the BMW Group” said Marjorie Janiewicz, chief revenue officer of Mistral AI. “This collaboration shows how industry specific AI models can help solve complex engineering challenges such as crash simulation.”

Large industry model as a technical foundation

To scale this approach, the BMW Group is focusing on large industry models. These are AI systems trained on industry-specific engineering and simulation data from vehicle development and safety testing. Unlike general‑purpose AI systems, LIMs embed domain‑specific knowledge directly into the AI model. This requires not only industrial data, but also deep domain expertise and technical environments that enable AI systems to learn directly from BMW’s development processes.

Related news, Next-generation Helm.ai models deliver full-HD 360° synthetic driving environments

The foundation models address what is often referred to as the industry’s data wall, where the cost and effort of collecting real-world edge-case data can slow development. Traditional generative world models typically produce video at sub-HD or VGA resolutions (around 0.4MP per camera). In contrast, Helm.ai generates full-HD (2MP) output that matches the resolution of modern production vehicle cameras, helping to reduce the sim-to-real gap in training for Level 2 and Level 4 autonomous driving systems.

Scene transfer versus fully synthetic generation

Helm.ai’s platform provides auto makers with a pipeline for data augmentation and creation.

GenSim-3 (high-fidelity scene transfer) enables development teams to restylize real-world video synchronously across six-camera, 360° surround-view setups. The model alters parameters such as weather, illumination and object appearance at full-HD (2MP) resolution. Additionally, the latest model introduces improvements in environmental texture, surface reflectivity and light behavior on complex materials.

VidGen-3 (fully synthetic generation) generates highly realistic driving sequences completely synthetically. By simulating complex environments, human-like agent behaviors and traffic logic from scratch, VidGen-3 bridges geographic and environmental data gaps at scale.

The 5X pixel density advantage

The technology’s key breakthrough is the fidelity of the multicamera generative simulation.

By producing full-HD (2MP) video, Helm.ai provides significantly more visual information than traditional generative datasets. Because modern production vehicles use high-resolution camera systems, training data is more effective when it matches that resolution. Lower-resolution synthetic data can create a domain gap when used to train full-HD perception systems. By generating data natively at 2MP per camera, Helm.ai aligns training inputs with real-world sensor output, supporting more consistent model performance in deployment.

To accommodate diverse sensor and training requirements, engineering teams can optimize for dynamic, high-speed validation with three-camera setups at 30fps, or maximize spatial context with a full six-camera, 12MP surround view at 5fps.

The virtual sensor twin

Unlike CGI-based video generation methods, Helm.ai’s models are designed to simulate hardware-like sensor output by incorporating certain physical characteristics of real camera systems. This includes reproducing effects such as sensor noise patterns, lens flares and exposure-related artifacts. This produces training data that more closely reflects real-world camera behavior, helping perception systems learn under conditions similar to those encountered in actual driving environments.

High fidelity on lower compute

While other generative world models rely on the massive computational scaling of thousands of GPUs to generate sub-HD video, the full-HD (2MP) resolution milestone was achieved using a highly optimized cluster of just a few hundred advanced GPUs.

“We are moving the industry from standard ‘AI video’ to authentic, hardware-accurate sensor emulation,” said Vladislav Voroninski, the CEO and founder of Helm.ai. “By leading with a full-HD (2MP) standard and a 12MP total aggregate capability per timestep, we have solved the resolution bottleneck that has historically limited the utility of generative AI in safety-critical systems. By optimizing our compute architecture, we are giving our partners a high-performance platform to validate their autonomous stacks using synthetic data that perfectly matches the fidelity of their actual production sensors.”

In recent news, Candera unveils “smarter and faster” CGI Studio 3.16 HMI development software

“Our customers expect a straightforward path from simulation model to microcontroller, and the new integration with Matlab and Simulink delivers exactly that,” said Brad Rex, senior director of the system solution team, user experience group at Renesas. “By working with MathWorks, we’ve removed the need to assemble toolchains and device drivers by hand so teams can simulate and validate designs earlier, iterate faster and reduce integration effort across ECU and industrial‑control projects.”

The Renesas RH850/U2A microcontroller – widely used in automotive ECUs – provides the deterministic performance and safety-critical features required for EV motor control, ADAS and body electronics. Engineers developing traction motor control for electric vehicles can deploy field‑oriented control and regenerative braking algorithms directly from Simulink to RH850/U2A‑based ECUs. This shortens the time from concept to vehicle‑level testing, supports smoother torque delivery during rapid transients and speeds calibration across drive cycles – without writing initialization code or custom build scripts.

Said Anuja Apte, India product marketing manager at MathWorks, “Our collaboration with Renesas strengthens the level of interoperability that engineers expect when using Matlab and Simulink. By providing a direct path from Simulink models to optimized microcontroller deployment, we help engineering teams move from design to hardware more efficiently while staying integrated with the broader toolchains they rely on. This approach reflects the MathWorks Connections program, which brings partners and customers together to accelerate innovation and reduce time-to-market within a widely adopted engineering and scientific platform.”

In related news, OmniTrust and Synopsys collaboration enables earlier security validation of embedded systems

The release extends the capabilities of Candera CGI Studio across the entire workflow from design and validation to deployment, and enables development teams to create and scale modern user interface concepts more efficiently across industries, hardware platforms and a wide range of use cases.

“With Candera CGI Studio 3.16, HMI development gets smarter and faster, combining AI-assisted tools, flexible workflows, and broad platform support,” said Roland Winkler, senior product development manager at Candera.

Improvements and updates

The Scene Composer experience has been enhanced with a new View Editor that simplifies automotive HMI workflows by allowing developers to preview scene combinations and manage transitions within a structured framework.

An Embedded Player enables seamless switching between design and runtime simulation inside Scene Composer, while a standalone Player remains available.

Multi View in the Scene Editor adds up to four synchronized views with selectable camera angles, improving inspection of complex 3D interfaces and enabling quick switching between single and multi-view modes.

Localization and documentation workflows are improved with built-in XLIFF 2.0 support. Developers can import, merge, edit and export localized text in a standard XML-based format, making CGI Studio easier to integrate into existing localization toolchains.

The new AI Docs Agent provides in-workflow support by delivering tailored answers in any language, helping users find information more quickly and reducing day-to-day friction. Each response also includes direct links to relevant Candera CGI Studio documentation, enabling users to access source material immediately.

Solution templates and advanced 3D

Another highlight of Candera CGI Studio 3.16 is the expansion of ready-to-use automotive HMI templates. The new In-Vehicle Infotainment (IVI) template includes building blocks for navigation, weather, media, phone, HVAC and digital twin scenarios. It supports both 2D and 3D use cases, with advanced visualizations and customizable application screens.

The new Instrument Cluster Solution Template integrates advanced driver assistant system functions such as adaptive cruise control (ACC), lane departure detection (LDD) and side assist alongside essential dashboard elements like speed gauge, media, telltales, time, odometer, temperature and turn-by-turn guidance. These templates help teams accelerate prototyping and shorten time to value.

The update enhances 3D capabilities with full glTF 2.0 support and features such as clear coat rendering. It also adds HDR rendering with tone mapping and automated import of image-based lighting cube maps for global illumination. Together, these improvements enable more realistic HMI visuals across automotive, industrial and embedded applications.

Engine and platform improvements

On the platform side, Candera CGI Studio 3.16 expands deployment flexibility with native rendering support for Apple iOS, enabling developers to build applications for Apple devices using Xcode and Metal.

The release also introduces Vulkan support as a modern graphics backend for 3D GPUs, improving CPU and GPU efficiency on Linux and Android platforms.

Another key addition is Software Rendering, which enables Candera CGI Studio 3.16 to run on platforms without a GPU. This enables platform-independent rendering across a wide range of targets, including ESP32-S3, ESP32-P4, Traveo II Body MCU, Raspberry Pi Pico and Linux-based systems. The renderer supports transformations, blending and visual effects while achieving up to 60 FPS depending on the platform and configuration, with low memory requirements.

Additional engine optimizations round out the release. Candera CGI Studio 3.16 now supports the new Monotype Spark Engine, designed for low memory usage and high performance. The update also includes newer versions of FreeType and HarfBuzz, delivering significant cache improvements for more efficient text rendering.

Candera CGI Studio 3.16 also streamlines behavior configuration through automatic code-size optimization using the SC plugin and CMake integration. These improvements help developers reduce memory footprint, improve performance and simplify overall project setup and configuration.

Recent news, Michelin reveals software-based tire digital twin platform

“It was a great start yesterday (May 19) with the launch of our smart prototyping concept with lots of positive feedback from customers and partners,” said Tanneke Reinders, executive VP at HBK. “From this morning’s presentations we really see ‘smart testing’ coming alive. The key is transforming from data creators to data interpreters – that’s the main shift for us. It’s about connecting data in one ecosystem. When we talk about data, especially the amount of data and fragmentation of it, we want to support.”

This year there were multiple tire-focused presentations, highlighting the critical role of the tire and its relationship with the vehicle. It seems tire and vehicle developers are working closer together than ever to refine their use of simulation and tighten the integration between the digital and real worlds. This was exemplified in the opening presentation by Alessandro Capobianco, R&D specialist for tire models and vehicle/DIL simulation at Bridgestone, and Eric Walter, vehicle dynamics CAE engineer at VW (both above), who provided an insight into their vehicle/tire co-development work.

Another of ATTI’s highlights from the summit was the Nissan and Horiba Mira presentation (right), in which the two speakers, Leo May, lead simulation engineer, Nissan Motor Corp, and Stratos Stratoudakis, a consultant at Horiba Mira, detailed how they are working together to prioritize virtual development by moving it further forward in the cycle, with the ultimate aim of reducing development time by 25%.

Marc Hedrich, a development engineer at event exhibitor Team Rosberg Engineering, said of the company’s attendance at the summit: “We are presenting our expertise in providing simulation models. The focus for us is to learn new things because software is ever evolving and you don’t want to be left behind. I like this year’s approach of trying to get the best data and not the most. I’ve learned already how others such as Mercedes-AMG are using these tools. It’s cool to connect with companies that are doing similar things and others that are doing things we’ve never heard of.”

This morning concluded with a presentation from the rarely heard-from Tesla (top), which gave insights into the company’s NVH analysis using a simulator.

Francesco Calabrese, who was attending as an exhibitor representing Fraunhofer ITWM as well as the institute’s spin-off, Virtual Tire Technologies, said, “It’s a good event to push the word for using more simulation with the goal of spending less money on prototypes. What I like is the excitement in this field as it grows – every year there are significantly more people which is a sign people are understanding the importance of it.”

Another OEM getting to grips with driver-in-the-loop simulation was Alpine. Rémi Auman, expert leader in simulation for ride and handling, noted that his team is targeting three areas for 2026 (see image below): enhancing driver immersion – an ongoing challenge for all developers; increasing the use of sim-workbench services; and improving multibody models. For comfort studies, Auman stated that the auto maker’s models are quite mature, so the next steps will be to adjust the settings and, “maybe in the future, we’ll be able to achieve real-time simulation,” he said.

With the conference’s emphasis on data harnessing, Cosin Scientific Software, which has been a partner of VI-grade for more than 15 years, launched its fresh approach to tire data management. With an increasing amount of users, the company has run into scaling problems when distributing high-quality input data. Developing countries especially cannot afford high-quality input data. To address this, the company has created a marketplace with a digital rights management system where tire data providers can market their data through Cosin’s software on the platform.

Cosin MD Gerald Hofmann explained, “Up until now, tire data has been sold as data files. This means once the data is out your hands, then it’s out in the wild. You can’t build a business model on this, which is why this data is very expensive. By putting the data under a digital rights management scheme, then the data provider can sell it using pay-per-use models.”

According to Hofmann, the development of this tire data library was an unexpected but welcome development for visitors: “We are very excited to see how this will be adopted through workflows which are currently very static. We’re hoping to break this up by giving access to a variety of input data. Users can allocate budget for a library of tires to choose from during the development cycle.”

AI specialist JuliaHub exhibited its latest technology, Dyad 3.0, which combines scientific machine learning with Gen AI. According to JuliaHub head of business development Europe Michael Hoffman, its name comes from its dual capabilities in performing controls and system modeling.

Asked what the company’s USP is, Hoffmann emphasized that having a single AI agent that can replace multiple tools is a powerful offering: “When we started Dyad we had AI mind – it wasn’t an afterthought where we had to do a plug-in. This means that the agent knows our architecture, and it only has to deal with one language.”

As expected, Dyad was met with both excitement and skepticism at the HBK SPS.

“We feel when talking about smart testing, we can add value,” added Hoffmann. “You have the measurements and we have the people to get the mathematical models or to refine their simulation models. For example, our value proposition to Rebel Dynamics [a servo platform partner of VI-grade], is that they have a mathematical model of their system, but they also have measurements, and there are differences between the two. We could help them minimize those differences so they can achieve better control.”

Related news, rFpro develops Japanese test route digital twin for virtual vehicle development

“Vehicle manufacturers are being asked to bring new models to market faster, at lower cost and with more variants than ever, and the only way to do that is to move more of their development work into simulation,” said Catherine Wood, head of content at rFpro. “That only works if the digital environments, which engineers are relying on, are highly accurate and correlate to real development decisions. The Hakone model is proving popular with our customers as it’s a route their test drivers already know intimately.”

A route built at scale

The digital twin has been created using survey-grade lidar scan data to create a vehicle dynamics-grade road surface that is accurate to within 1mm in height across the entire 15km route. Every curb, drain, paint marking and barrier has been placed accurately to match its real-world counterpart. Where relevant, the paint markings are worn and cracked to increase realism and immersion for the driver.

Building a route of this nature and scale requires a careful balance of visual detail. A driver-in-the-loop (DIL) simulator has to run in real time, so rFpro’s content team created engineering-grade fidelity where the driver and the vehicle model need it most. Detailed modeling extends approximately 25m either side of the road, where accuracy is critical for vehicle dynamics and where the driver’s eye is primarily focused. Beyond that, the surrounding landscape, which includes the distant hills and Mount Fuji, is represented by an optimized 3D model of the terrain, incorporating satellite imagery. This helps to balance run-time performance and driver immersion.

The Hakone Turnpike model includes 57,000 trees, 13km of Armco barriers, 460 drains and more than 45km of painted road markings. The tollgate and plazas at the entrance are fully modeled. The route also features 12m-high retaining walls that were particularly challenging to replicate. They sit directly in the driver’s field of view and any obvious repetition in their brickwork texture would be immediately noticeable, so the team used rFpro’s advanced material shaders and techniques to produce long, unique sections that look highly realistic even at close range.

Expanding Japanese portfolio

Hakone Turnpike joins rFpro’s growing library of Japanese content, which includes public road models of Tokyo’s C1 urban expressway, the Tomei expressway and numerous race and testing circuits, such as Suzuka, Fuji, Okayama, Motegi, Sugo and Autopolis.

rFpro has also completed high-fidelity lidar scans of public roads in Odawara, Yokkaichi and Kyoto, and is working with customers to develop these into future additions to its public road library.

“Hakone is one of the most demanding builds we have taken on,” added Wood. “The scale, the vegetation, the walls and the sheer quantity of surface detail all had to be handled without compromising simulator performance. The result is a route our customers can use with confidence for vehicle dynamics, and is incredibly immersive for drivers who already know it well.”

In related news, Daimler Truck to use VI-grade simulator to enhance ride comfort and vehicle development

The simulator is specifically designed to support detailed subjective assessments of ride comfort across different vehicle concepts. It enables evaluations at up to two seating positions and reproduces realistic vibration behavior, giving engineering teams a better understanding of vehicle dynamics and comfort perception early in the development process.

The equipment integrates advanced measurement technology, enabling the simultaneous acquisition of test bench data and simulator inputs. It supports a wide range of excitation sources, including simulation-based models and real-world vehicle data, ensuring high accuracy and repeatability. To complement these capabilities, HBK’s nCode software is used for advanced data processing and analysis. It enables engineers to efficiently process measured data, perform durability and FE-based fatigue analysis and gain deeper insights into vehicle performance, while supporting robust data management across the development workflow.

“Ride comfort and vibration development in commercial vehicles comes with a high level of complexity, driven by diverse operating conditions and vehicle configurations,” said Helmut Schmitz, Central Europe sales director of simulation at VI-grade. “Our simulation solutions enable engineering teams to replicate these conditions realistically, combining subjective evaluation with precise data to accelerate development and improve overall vehicle performance.”

In related news, BMW M Motorsport partners with MdynamiX to advance virtual driving for vehicle development

BMW M Motorsport will use MdynamiX’s extensive expertise in driving experience evaluation, virtual development methods and precise steering system simulation. MdynamiX will also provide its proprietary high-performance force feedback steering systems for this purpose.

The technology is designed to deliver a highly realistic driving experience through the steering system, which serves as a key interface between the driver and the vehicle.

Using proprietary force-feedback hardware and detailed steering system simulation models, racing drivers and engineers can evaluate steering characteristics early in the development process, allowing vehicle dynamics and steering feedback to be optimized before physical prototypes are built. The development environment enables complex steering systems to be virtually validated under racing conditions, with faster iterations and reliable decision-making foundations reportedly resulting in improved performance and shortened development cycles.

“The importance of virtual driving experience development in motorsport continues to grow,” said Erik Schuivens, head of vehicle performance and simulation at BMW M Motorsport. “With MdynamiX, we are working with a partner that not only provides methodological support but also delivers the technological foundation through its proprietary force feedback steering systems, allowing vehicle behavior to be experienced with exceptional precision in the simulator.”

“BMW M Motorsport stands for precise engineering and performance,” added Prof. Dr Peter Pfeffer, the CEO of MdynamiX. “The fact that our expertise in steering system simulation and our force feedback hardware now contribute to the performance of these vehicles validates our approach. Together, we are setting new standards in virtual driving experience development.”

In related news, Ansible launches Delta T1 Sport simulator

The first production unit will be delivered to Lola Cars to support its vehicle development programs, including preparations for its ABB FIA Formula E World Championship GEN4 campaign.

Optimized for single-seater and cockpit racing series, such as Super Formula, LMP2, LMH/LMDh, Formula E, F2, F3, F4 and Formula Regional categories, the Delta T1 Sport reportedly opens the door to engineering-class DIL capability for race teams, drivers and organizations where space, infrastructure, installation complexities and cost of ownership have traditionally been barriers.

The simulator features an all-new Triform motion system that blends a uniquely high mechanical stiffness and low dynamic mass, to deliver elite physics, along with trusted, repeatable, tunable motion performance, Ansible reports.

Dan Clark, Ansible Motion’s managing director, said, “Delta T1 Sport delivers the fidelity that drivers and engineers crave, in a compact, easy-to-install and operate, inclusive package that has not been available until now. This new class of simulator will unlock fresh opportunities for teams to efficiently pursue and extract vehicle performance gains by leveraging a streamlined, optimized approach. Our customers will benefit from best-in-class, combined solutions in terms of technical integration, technical support and value, with the freedom to distinctively configure aspects of their DIL simulation environment that ultimately lead to competitive advantages.”

Delta T1 Sport in Formula E

The first production Delta T1 Sport will be delivered to Lola Cars’ Silverstone headquarters. The British motorsport company will draw on the benefits of virtual development for its Formula E program, with a wider range of DIL-simulator-supported initiatives set to follow in the longer term.

Till Bechtolsheimer, chairman of Lola Cars, commented, “We are delighted that Lola will be the first user of the pioneering Delta T1 Sport simulator, which will efficiently connect to our existing workflows, enabling us to accelerate our engineering processes and improve development as we strive to drive innovation through motorsport.”

Peter McCool, technical director of Lola Cars, continued, “We do not underestimate the impact this technology will have on our car development, software validation and race preparation, for both Formula E and future projects.”

Scalable architecture

As the latest addition to Ansible Motion’s DIL Sport simulator portfolio, Delta T1 Sport follows the Theta Seat Sport (already trusted in Formula E) and Theta Cube Sport (used by a European motorsport customer).

The DIL Sport range is designed for motorsport engineering applications, offering high dynamic fidelity, ultra-low latency and enhanced driver realism. The system also supports seamless integration with racing software and hardware environments to help in vehicle development and testing workflows.

“Within six-degrees-of-freedom motion space, DIL simulators must, first and foremost, deliver sufficient ground-plane fidelity – especially for vehicle dynamics and motorsport applications, where vehicle directional control and stability dominate the virtual test-driving experience,” said Elliot Dason-Barber, Ansible Motion’s technical director. “Many simulators miss this fundamental point by starting with legacy hexapods and then working backward to sort out additional motion requirements with appendages that add complexity and cost.”

Compact, six-degrees-of-freedom motion system

At the core of the Delta T1 Sport is Ansible Motion’s Triform motion system (AML TMS1), which is designed for motorsport and high-performance simulation applications.

The system uses updated kinematics and motion generation technology, aiming to retain elements of the company’s Stratiform motion approach while reducing the overall physical footprint.

The Delta T1 Sport offers a payload capacity of up to 300kg, supporting drivers and racing cockpits, and provides a six-degrees-of-freedom motion system with 1m surge, 1.2m sway, 0.2m heave, 22° roll, 16° pitch and 120° yaw.

Ansible Motion says the motion envelope has been optimized based on its experience in DIL simulation, to support realistic feedback on vehicle dynamics, including handling, track conditions and stability.

The system also includes a configurable motion cueing package, allowing sensory feedback to be tailored to different racing series, driver profiles and vehicle setups.

This overall motion space is efficiently packaged within a low-mass, 2.4 x 2.4m floor footprint, which enables flexibility of location and operation without special facility or flooring requirements.

Salman Safdar, Ansible Motion’s business development director, said, “The Delta T1 Sport also includes a number of bespoke driver feedback and immersion systems to suit physical space, budget and performance requirements, as well as futureproofing and upgrade pathways. These include, but are not limited to, multiple space-efficient vision systems, including projectors, LED walls, extended reality (XR) and virtual reality (VR) options; configurable race car cockpits; active dashboard and steering wheel interaction capabilities; adjustable racing seats; configurable pedals; haptic and torque feedback for steering; and harness and helmet loading systems.”

In related news, AVL and Ansible advance vehicle validation with integrated simulation and driver-in-the-loop testing

Automotive development is evolving as software-defined vehicle programs introduce faster feature cycles and more complex system interactions while meeting strict requirements for safety, reliability and long-term maintainability. Gen AI is now part of engineering workflows. It can help increase development speed, but its non-deterministic behavior, lack of physics awareness and limited traceability make it difficult to apply directly to safety-critical systems. These characteristics make verification, certification and traceability challenging when outputs generated by Gen AI are introduced without constraints.

Model-based design addresses these issues through deterministic execution, executable specifications and physics-based simulation. MathWorks is bringing these strengths together by integrating Gen AI assistance directly into model-based design tooling, enabling engineers to benefit from accelerated workflows while preserving the rigor required for long-term reliability and certification of automotive software.

Simulation as the foundation of trust

Simulation is the foundation of trust in engineering workflows assisted by Gen AI. It provides a controlled environment where system behavior can be verified early and repeatedly. Model‑based design enables closed‑loop simulation within continuous development pipelines, enabling Gen AI‑assisted artifacts to be validated continuously in virtual environments long before software reaches hardware. Closed-loop simulation uncovers defects that emerge only from real‑time interaction between software, hardware and physical dynamics, such as instability, timing issues, saturation and integration errors. Unlike regular software tests that validate code logic in isolation, simulation validates system behavior against requirements under realistic operating conditions, catching safety‑ and performance‑critical issues much earlier.

In leading organizations, ‘shift left’ is not a one-time activity; virtual verification is embedded directly into continuous integration/continuous development (CI/CD) pipelines. Every change triggers automated builds and simulation runs, exercising models against representative scenarios and acceptance criteria. Verification becomes continuous, not episodic.

Scalable development for evolving E/E architectures

Automotive E/E architectures are transitioning from ECU-centric networks to zonal and centralized computing platforms. Software is no longer bound to specific hardware configurations but must now operate reliably across heterogeneous compute targets while remaining portable and scalable, from small controllers to high-performance vehicle computers.

Model-based design supports this requirement by separating system behavior and software intent from hardware implementation. Engineers develop executable models that serve as stable sources of truth. The models can generate production-ready code for a wide range of processors and operating systems, including platforms incorporating AI inference engines and hardware accelerators such as GPUs, DSPs and NPUs. This approach enables the development and validation of AI-enabled functions (e.g. virtual sensors) at the system level, reduces the need to reengineer algorithms for each target, and improves efficiency and consistency across platforms.

Improving collaboration through model-based design

Engineering organizations must transform their collaboration models to keep pace with increased complexity. Integrating simulation, virtualization and automated verification directly into CI/CD workflows supports rapid iteration across software, AI models and hardware acceleration strategies. This model-centric approach helps organizations operate more quickly while preserving robustness, safety and long-term maintainability in the era of software-defined and AI-driven vehicles.

Integrating AI into deterministic engineering workflows

AI is most effective in automotive development when embedded within a deterministic modeling framework. Within model-based design tools, GenAI-generated content is automatically tied to established interfaces, data definitions and architectural constraints. Model Context Protocol (MCP) capabilities empower engineers with AI assistance while preserving the rigor, repeatability and certification readiness.

Long-term maintainability and certification readiness require deterministic behavior, transparent audit trails and verification evidence that accumulates throughout the lifecycle. Model-based design naturally supports these goals by linking requirements, models, test suites and generated code. Continuous simulation produces verification data throughout development rather than only at the end of a program. When artifacts generated by Gen AI follow the same workflows, they inherit this structure. This ensures that productivity gains do not come at the cost of safety, quality or compliance, and that Gen AI can be adopted at scale.

Conclusion

Gen AI and model-based design offer a structured path to accelerate automotive software development while maintaining trust, safety and engineering rigor. Model-based design provides determinism, physics-based validation and traceability. Gen AI adds efficiency and supports faster iteration when integrated within these boundaries.

This combination enables earlier insight into system behavior and deployment across diverse hardware architectures. The model-centric approach ensures consistent collaboration across engineering teams, and promotes reuse and consistency across global programs. Gen AI-enabled model-based design provides a scalable and reliable foundation for developing robust and certifiable automotive systems.

In the September 2024 edition of ATTI, Secondmind’s chief product officer, Morgan Jenkins, discusses the power and limitations of AI

In related news, Agentic AI transforms McLaren Automotive’s entire engineering process