Automotive and AI – where is the journey headed? The IT service provider adesso explains which AI issues should be a key focus for the automotive industry:
AI in business processes
AI has already become an integral part of modern vehicles in driver assistance systems, route optimisation and voice-controlled radios. However, the automotive industry’s business processes are a different story. Although the industry belongs to the high-tech segment, AI still does not see widespread use. Established companies in particular are finding it hard to transform themselves from traditional automotive manufacturers into data- and software-driven mobility service providers. They need to focus more strongly on business processes and applications for AI outside of the vehicle. As a motivating impulse, adesso recommends taking these points to heart:
- Give AI a face and make it a tangible experience: application scenarios have to be identified and tested in order to get employees and the various departments and IT ar-eas excited about AI.
- Give AI a home: establish central platforms and competence centres as service units that provide expertise on demand.
- Put AI where the money is: test AI business cases, evaluate them critically and then follow up and communicate the top cases.
AI in production
AI supports production through image, data and sound pattern recognition for monitoring ro-bots, systems, transport routes and quality assurance. Trained AI systems can discover even the finest cracks in a windscreen during a visual inspection or anomalous engine noises during an acoustic test. By taking steps like these, manufacturers and suppliers can optimise their production and logistics processes, immediately identify quality deficits and prevent sub-sequent problems. Thanks to advanced knowledge in the field of deep learning, such scenari-os are realistically achievable today – with low error rates and high levels of maturity.
AI in the supply chain
AI independently interprets problems and develops suitable solutions. AI-based algorithms do not utilise a rigid framework. Instead, they verify the results based on a constantly growing database and are constantly improving. This reinforcement learning principle enables ma-chines to make increasingly accurate predictions. Logistics is one example: Based on large quantities of data that have an influence on the delivery of parts for vehicle production, AI ap-plications independently determine the best transport routes. In the age of just-in-time delivery throughout global supply chains, this represents a valuable application. Over the course of the coronavirus pandemic, the industry has been confronted with just how sensitive the supply chain is and how drastic the impact of shortages can be.
AI for drivers
AI-based personal assistants take the user interaction between drivers and vehicles to an en-tirely new level. Gestures and facial recognition, alongside spoken commands, play a major role. For example, they enable contactless control of infotainment systems or can inde-pendently unlock a vehicle. Integrating smart watches enables a vehicle to identify the driver’s emotions. As a result, the driving experience can be adapted to the driver’s emotional state and emergency situations identified earlier – a further advancement in safety for vehicle driv-ers.
AI and intelligent mobility services
AI makes intermodal mobility scalable. It optimises the usage and provision of the services which simplify connecting diverse means of transport such as vehicles and public transport while paving the way towards “mobility from a single source”. In the process, intelligent pricing models manage supply and demand. AI places the focus on the customers when it comes to communication, providing the service and offering support.
AI and autonomous driving
Autonomous driving is impossible without AI. AI systems in autonomous vehicles process a wide range of data simultaneously and in real time, such as the course of the road, road signs and traffic lights, the movements of other road users and much more. The advantages are clear: AI systems are capable of receiving, processing and interpreting vast amounts of data simultaneously. In addition, they are not distracted by smartphones, radios or eating while driving. As long as the power supply and the technology function properly, they never grow tired. Plus, the longer and the more often they are used to process new data, the better and more precise they become.
Although the breakthrough of self-driving cars on our roads will not happen tomorrow, re-search and development continue to advance. AI systems have to learn to interpret the data. The developers utilise extremely comprehensive datasets to train the AIs and test their deci-sions. As the tests progress, the AI system develops the ability to navigate safely on the road. It learns to rapidly react to unforeseen events such as an abruptly braking vehicle. Expanding sensor systems beyond the vehicle to encompass the cyber-physical environment – such as traffic-management systems, parking guidance systems, smartphones and other means of transport – creates both safer route guidance and more efficient routes.
