Frugal innovation through applying AI to existing data
Innovation is critical to dealing with industry changes, but innovation does not necessarily mean invention; often, it’s about reusing an idea or a resource in a different context. This frugal approach to innovation, called ”Jugaad” in India, is a theme of Capgemini’s TechnoVision for Automotive 2022 playbook. The application of AI is a prime example of how the automotive industry is innovating by making better use of what it has. Of course, AI needs data, and this is where the frugality comes in, because much of the necessary data often exists already. For example, data generated by designing and building a vehicle was often discarded after the product’s completion, but with software increasingly determining which options an individual vehicle offers, its value is clear. AI can help companies make the most of data in a range of contexts.
Inside the vehicle: Perhaps the best-known use of AI in automotive is to automate the task of driving. Even if fully autonomous vehicles are still a few years away, Advanced Driver-Assistance Systems (ADAS) features are already appearing. AI can open up a whole world of seamless driver interactions and can support safety, reliability, and robustness.
On the factory floor: AI is integral to many current production-line automation initiatives designed to increase efficiency and quality. AI-based systems can help to analyze camera outputs, carry out shop-floor quality checks on the assembly line, optimize truck loading to improve space utilization, or power augmented-reality goggles to minimize operator errors. In the design workshop: With AI, revolutionary propositions can emerge from data, including new elements for use by human designers. AI can also evaluate solutions generated by humans or machines and recommend the most promising.
In the back office: For strategic planning purposes, AI-enabled processes can assist with rationalizing the choice of vehicle configurations. AI could even help the human resources function because, when talent is scarce, AI can make the most of the people you have.
AI helps optimize ADAS General Motors is assessing the potential of an AI-enabled pattern recognition technology to accelerate the design of ADAS. The Multi-node Evolutionary Neural Networks for Deep Learning rapidly evaluates convolutional neural networks for use in pattern recognition. This approach could, for instance, reveal ways for cars to quickly and accurately assess their surroundings in order to navigate safely through them.
Sharing safety data Capgemini has been working with Volkswagen and Audi to demonstrate the value of the German Federal Government’s Mobility Data Space, of which Volkswagen Group is a founding member. An early use case is Local Hazard Information, which provides aggregated event data on traffic hazards collected from vehicle sensors in the Audi fleet. This data could be used by a navigation service to warn road users of upcoming danger spots in near real-time.
