One car can hide another
Despite appearances, some of the vehicles on the streets of San Francisco, Phoenix in Arizona or even Beijing do not always have a driver at the wheel: these cities were among the first to authorise the use of robot taxis. These are autonomous vehicles that are fully automated and driver-less. Such vehicles have reached level 5, the highest level of autonomy (see insert), thanks to sensors and artificial intelligence that can detect the environment, combine the information obtained to analyse it, decide on an action and implement it.
Europe classifies its cars in five levels of autonomy.
At levels 1 and 2, the on-board software brakes, accelerates or keeps the car in its lane, but it merely assists the driver, who remains solely responsible.
True autonomy begins at level 3, where part of the responsibility is entrusted to the software (and therefore indirectly to the manufacturer), for overtaking a car on the motorway, for example. The driver must nevertheless be ready to regain control in the event of an alert from the vehicle, and this is always the case at level 4. At level 5, the driver is no longer needed and the car is fully autonomous.
Honda is the first manufacturer to market a Level 3 partially autonomous driving system in Japan in 2021, followed by Mercedes-Benz in Germany in the same year, and in Nevada and California in 2023 - and soon BMW in Germany this year.
While several operators, such as Waymo (Google / Alphabet), Zoox (Amazon) and Nuro have taken the plunge, most carmakers are marketing level 1, 2 or even 3 autonomous vehicles using artificial intelligence. The benefits for the automotive industry and the user experience are already immense. Since September 2022, Germany, for example, has authorised Level 3 "autonomous driving vehicles", which can drive semi-autonomously on motorways, in traffic jams and in car parks. Following in the footsteps of Honda and its Legend Hybrid EX, authorised in Japan in 2011, Germany's Mercedes-Benz has been authorised to market its Level 3 models in Germany, China and California.
At present, Level 2 and Level 3 partially autonomous cars are the most used on the market, while Levels 4 and 5 (as scaled by SAE International) are expected to become more widely accepted by 2030.
(Source : mordorintelligence.com, autonomous-driverless-cars-market-potential-estimation)
AI for safety
Thanks to artificial intelligence, connected vehicles, even if they are not 100% autonomous, can already significantly improve on-board safety. AI has a formidable capacity for absorbing and exploiting data from both inside and outside the vehicle. It provides permanent surveillance solutions, using sensors and cameras to assist the driver. And AI is at the heart of advanced driver assistance systems (ADAS): automatic braking, collision warning, speed regulation, lane departure detection, etc. Most manufacturers have also introduced fatigue detectors in their vehicles, which analyse driving behaviour and warn the driver at the first sign of a loss of concentration.
Being able to anticipate what is happening in and around the vehicle, with techniques such as emergency braking, can significantly reduce accidents, which is what carmakers are looking for. All these technologies are increasingly being deployed, including in standard vehicles.
Matthieu Soulé, Director of BNP Paribas C. Lab Americas
Optimised fleet management
And that's not all. AI also plays an important role in predictive maintenance. Thanks to the real-time analysis of this on-board technology, AI can anticipate mechanical failures by detecting variations in temperature or pressure, for example, and suggest the replacement of certain components. By keeping the vehicle in optimum condition, AI helps to enhance on-board safety.
Predictive maintenance is becoming an essential part of fleet management. AI makes it possible to process a multitude of data on vehicle components - and their potential obsolescence - but also on their general state (vehicle performance, efficiency or productivity). "By processing millions of data items using AI fleet managers have a source for cost optimisation and useful information on consumption, usage, etc.", analyses Matthieu Soulé.
BNP Paribas Arval, one of the leading operational leasing companies and a specialist in mobility solutions, has just passed the 100,000-mark for connected vehicles thanks to data received directly from vehicle manufacturers. These vehicles are among the 650,000 connected vehicles in Arval's fleet thanks to the Arval Connect solution. The BNP Paribas subsidiary aims to connect 80% of its fleet by 2025. This is a way of organising proactive maintenance operations and optimising fleet management by reducing downtime.
Volkswagen has announced its ambition to spend more than 122 billion euros on electrification and digitalisation between 2023 and 2027; in 2023, Honda and Sony launched Afeela, a mobility brand with an ultra-connected zero-emissions prototype, equipped with 45 sensors - LiDARs (Light detection and ranging), cameras, etc. - distributed around the outside and inside the vehicle. Ultimately, the prototype offers a digital platform that provides driving aids, semi-autonomous piloting and connections to different networks (between vehicles, between vehicles and infrastructure, etc.) via 5G in particular.
More recently, French carmaker Stellantis announced a partnership with Amazon to design the software for its "smart cockpit", an electronic platform that will be integrated into the vehicles of the group's 14 brands. This is an extension of the Stellantis software strategy set out in the Dare Forward 2030 plan, which aims to develop intelligent mobility products and enhance the customer experience through personalised functionalities.
All carmakers are making massive investments in software, and partnerships between manufacturers and tech companies are now commonplace,"
Guillaume Rio, Technology trends and partnerships manager, BNP Paribas BCEF
AI and sustainability: a long-term relationship?
The impact of AI on the automobile goes beyond safety and convenience; it is also shaping the future of sustainable mobility. AI can be used to optimise journeys made by connected vehicles, helping to improve traffic flow, control the filling of autonomous vehicles, provide advice on eco-driving, parking and relaxation, etc. By dynamically adjusting parameters, on-board software measures and advises on how to reduce fuel consumption and CO2 emissions.
From the driving experience to improving on-board safety and optimising journeys, AI is far from finished taking the automotive industry into new worlds.
