While the EV ecosystem may get most of the mainstream press, another simultaneous evolution is reshaping the auto industry: The rise of the software-defined vehicle (SDV). Today’s automotive innovation – EV and otherwise – is all about software, and the AI boom is further accelerating the pace of change. In a conversation with Engineering Edge, Thomas Mueller – VP and CTO, Wipro Engineering and Dr. Swarup Mandal, Global Head Automotive, Wipro Engineering Edge discussed how AI will impact automotive engineering today and in the future.

Q: What’s the most important mindset shift automobile manufacturers will have to make with regard to AI?         

Thomas: If they haven’t already, automakers need to recognize that there has been a fundamental change. Automobiles were once a completely analog product. They are now connected smart devices on wheels. The industry needs to think of automobiles as durable digital-enabled devices that can be updated for at least 10 years.

Manufacturers will increasingly differentiate themselves through software. By creating digitally "evergreen" cars, they can enable continuous innovation — vehicles that get better and smarter after they leave the dealership. They will also gain more flexibility to address the unique needs of different owners. This software-defined innovation will be forcefully accelerated by AI.

Thomas Mueller

VP and CTO, Wipro Engineering

Any opportunity for software-defined differentiation will be touched by AI, and often by GenAI

Just one small example: AI has great potential to optimize cabin comfort and energy efficiency by leveraging surface heating and cooling systems. Use cases like this will be repeated across electrification, battery management, data analysis, autonomous driving, predictive maintenance, and much more. Any opportunity for software-defined differentiation will be touched by AI, and often by GenAI.

Dr. Swarup: Nowadays, when we engage in customer conversations, we are actively considering GenAI. The major bottleneck of using AI-enabled inference techniques within vehicles comes from the need for large data processing. With the advent of GenAI, the inferencing capability is enhanced and eliminates the need for processing large amounts of data. We are seeing increasing interest in including GenAI-enabled features. In addition to the above, we are also considering its role during the development life cycle, such as automated code generation, code performance optimization, and compliance. For instance, we have encountered situations where an automated test suite provides

100% success, but when it undergoes further testing in the field, we discover errors or defects at the end of the development lifecycle. To overcome this, we are exploring two approaches. First, we are using GenAI-based tools to generate corner scenarios and cover most of the testing requirements. Second, when errors occur, we are employing the explainability features of GenAI, which help to identify the root cause of the issue and thereby generate additional test cases to enhance the test suite.

Q: What are some applications of GenAI in automotive?

Thomas: We know that AI has the potential to help build cars faster and better. Cars now contain 100 million lines of code. AI can assist in managing the complexity of that code, and will also enable future software products to learn from user interactions in real time. The result will be an improved user experiences, and cars that can adapt to changing needs. To put it simply, with the right mindset shift, it’s possible to use AI to create better vehicles that cost less and have higher residual values. The manufacturers who do this first will have a serious advantage over those who delay.

Dr. Swarup: There are two broad categories of application for GenAI. The first category is GenAI-enabled features in the vehicle, such as voice assistants for reducing driver distraction, a traffic rule engine for assisting the drivers, and HD maps for autonomous driving, among others. The second category involves efficiently managing each phase of the vehicle's lifecycle, including design, development, manufacturing, and operation. The simulation capability of GenAI-based tools plays a pivotal role in this aspect.

Q: How does AI change the way automobile manufacturers will need to approach hardware and software? 

Muller: It changes it substantially. The concept of a "software-defined vehicle" (SDV) is crucial for the future of the automotive industry as it develops and mass-produces multiple vehicle models. Customer expectations have changed, and manufacturers need to keep up. AI will help.

Dr. Swarup Mandal

Global Head Automotive, Wipro Engineering Edge

The new paradigm in the automotive sector is the ‘AI-defined vehicle.’

An obvious and compelling need is to consolidate vehicle software to enable advanced driver assistance solutions (ADAS) and digital cockpits that operate on a high-performance computer (HPC). This allows for continuous delivery of new software through over-the-air updates. AI can learn how each specific driver in a household drives from day to day and adjust driving characteristics to enhance the driver experience, improve battery utilization, and more. Using AI, automakers will be able to optimize those ADAS solutions based on real-world data from daily use.

Dr. Swarup: The new paradigm in the automotive sector is the "AI-defined vehicle." SDV unleashes the potential of updating the vehicle, and AI brings in the capability to make this update contextual to the environment the vehicle is in and to the person(s) behind the wheel. OEMs will leverage more and more AI-enabled tools both in development and in operation. The targeted outcome defines the capability of the software to be supported by hardware. In other words, it is an "AI-first, software-driven approach."

Q: What are some other AI-driven features that drivers themselves will begin to notice?

Thomas: AI is being used to improve cockpit automation, automated parking, and lane keeping systems. Additionally, as Dr. Swarup already mentioned, GenAI plays a crucial role in the development of testing scenarios. That will ultimately have a positive impact on the driver experience. By generating new testing scenarios, GenAI helps ensure that automotive applications are thoroughly tested and bug-free.

Dr. Swarup: To elaborate on how AI will impact the user experience within the vehicle cabin: Companies like Ford are already working on concept cars that feature adaptive user interfaces. These interfaces adjust seat heights, inclinations, and dashboard layouts based on the context. For instance, if the driver is not actively driving while the vehicle is in a parked or autonomous mode, but rather engaging in a conversation, the seat may retract to create more space for work-related activities. This personalized and context-aware user experience enhances comfort and productivity.

Some companies have also developed rule engines that assist drivers in navigating different countries' driving rules. These systems provide advice and information on road conditions, blocked routes, and other relevant details. Additionally, projection lights are used during parking to show the trajectory and engage the surrounding drivers, making parking safer and more efficient. These real-world applications demonstrate the progress and potential of AI in the automotive industry.

Q: Where are the opportunities for profit?

Thomas: By creating a long-term and scalable vehicle computer platform, manufacturers can leverage the cloud to deliver new features and functionalities over time.

Thomas Mueller

VP and CTO, Wipro Engineering

Ensuring room for software growth over time is what enables vehicles to get better every day and every year. The industry has some unlearning and relearning to do. But it’s worth the effort.

Changing the core functionality of an engine is hard and would obviously require costly recalls. But changing the way the engine “thinks” is a different matter entirely. By leveraging AI, manufacturers can continuously roll out new infotainment features, safety enhancements, and core functionality updates over the air. We can expect some of those software updates to be free, but others can be optional upgrades that buyers can either buy as a one-time purchase or as a subscription. There’s good profit to be had there.

Dr. Swarup: AI will drive profitability along with affordability in three ways. First, it will make vehicle development and operation efficient and optimized, which will reduce the cost of development, focusing more on R&D cost optimization. Second, it will enable a secondary revenue stream by implementing a pay-per-use model for consuming the features. The third important potential lies in the large amount of data the vehicle captures while in operation. A significant revenue stream is yet to be tapped to add to the bottom line.

Q: Given what’s possible, why aren’t manufacturers doing more of this already?

Thomas: Change, especially in large-scale manufacturing, is always easier said than done. I totally understand there’s a learning curve for any company to handle the complexity of integrating AI in vehicles — and that’s just one of the ways Wipro can help. But it’s more than that. Manufacturers aren’t used to the idea of over-provisioning computer hardware in cars — it can feel like an unnecessary expense, especially if the driver won’t see the benefit of it on day one. Hardware changes and modifications are costly and time-consuming. But ensuring room for software growth over time is what enables vehicles to get better every day and every year. The industry has some unlearning and relearning to do. But it’s worth the effort.

Of course, that requires a different kind of infrastructure. That’s why at Wipro, we developed our Cloud Car ecosystem to empower OEMs to accelerate the development of software-defined vehicles. The API-ification of vehicles within the Cloud Car ecosystem enables developers to prototype new functions rapidly and for agile and responsive design approaches. This can help enable personalization features and dynamic customization within safety boundaries.

Dr. Swarup: Transforming the vehicle to embrace the new generation E/E architecture and deploying applications as microservices will enhance the reusability of subsystems of a vehicle platform. This can reduce the total cost of ownership for new or updated features by 30% to 50%. The use of microservices and containers allows for scalability as well. The continuous development and integration of software features can deliver services and user experiences that exceed customer expectations. All of these initiatives aim to make it easier and less risky for automotive manufacturers to embrace innovation. At Wipro, we are actively supporting major Tier 1 and OEMs in their transition to SDVs.

Q: What do automobile manufacturers have to do to get ready? 

Thomas: At minimum, we can expect AI adoption in the automotive industry to increase. As leaders embrace it, the rest of the industry will have to follow. Software-defined vehicles, continuous over-the-air updates, AI-driven predictive maintenance, real-time vehicle diagnostics, personalized experiences, and autonomous driving technologies will all become standard aspects of owning a car. 

I’ll go out on a limb and say that by 2028, most new passenger cars will be equipped with both automated driving features and AI-driven comfort featuresTraditional automakers need to shift towards a software-defined approach in response to the increasing complexity of modern vehicles. It’s time to decouple the hardware-software relationship and recognize the benefits of a cloud-native structure for automotive software development.

Dr. Swarup: Automobile manufacturers are increasingly interested in incorporating AI solutions into their digitalization processes. They need to collaborate with each other to reuse assets that are not differentiating for them, in order to save costs. The additional savings thus made can be used to build critical features or provision extra computing resources to better realize the value of SDVs. Each of them needs to traverse the journey of transformation at an accelerated pace and in a cost-effective way. This calls for a well-developed ecosystem in terms of software and AI capabilities to manage the development and operation of the vehicle. The OEMs should seek the help of engineering service providers like us, who are in a unique position to facilitate collaboration to advance SDV development with an AI-first, software-driven approach.

For instance, Denso, a leading automotive company, has expressed their desire to start small and gradually expand their AI capabilities to address specific pain points. More and more clients will take this route. Automakers understand the importance of maintaining traceability, especially for compliance. By combining systematic tools with human expertise, we can effectively overcome challenges and prepare for the future of AI in the automotive industry.

To learn more about Wipro Engineering services and how we use AI to enable our clients to realize their technology and business ambitions, please visit our engineering page and email us on info@wipro.com.