How can AI and generative artificial intelligence serve transportation?

What are AI and generative artificial intelligence?

Artificial intelligence may have only recently gained the media spotlight, but it has long been a part of the tech landscape. As early as the 1950s, researchers were exploring the possibility of replacing human intelligence with machines. Increased processor capacity, advances in algorithm research and the explosion of data volumes have enabled the development of machine learning and deep learning, the foundations of artificial intelligence. These computing techniques allow machines to learn by themselves and improve their performance in fields such as image recognition, text translation and obstacle detection. After just a few years, these techniques are opening up new horizons in mobility, notably with the rise of autonomous vehicles.

Generative artificial intelligence, born in 2017 and revealed to the general public in November 2022 with the launch of ChatGPT, represents a decisive turning point in the acceleration of artificial intelligence. It is based on enhancing existing data to produce new content (text, images, music, video, computer code, etc.). Today, its use has become more widespread thanks to software such as ChatGPT, Deepl, Perplexity and Mid Journey. 

Although the subject of AI is not new, its rapid development is having a significant impact on all business sectors, through a variety of applications. In particular, it is opening up new possibilities in the field of mobility. Keolis sees artificial intelligence as a fundamental evolution driving a series of major digital transformations. Like the Internet or social media, AI is profoundly transforming our lifestyles, and this is driving the Group to actively support its employees and passengers through these changes. The current acceleration of artificial intelligence, driven by the entire ecosystem (industries, startups, universities), is opening up new prospects for innovation.

Whether it is predicting ridership in public transit, optimizing preventive vehicle maintenance, calculating the CO2 trajectory of mobility across a given perimeter or improving safety, AI and generative artificial intelligence are now part of Keolis’ operations and services to improve the performance and quality of its networks. Let's take a closer look at three use cases.

Generative artificial intelligence in the service of passenger information

In the Île-de-France region, fares vary depending on the mode of transport and the lines used. For occasional users, especially tourists and visitors, this can lead to confusion and unintentional fraud. Keolis Île-de-France is therefore offering passengers a chatbot equipped with generative artificial intelligence, known as Keobot, which provides fare information in all languages. The service is accessible via a QR code that redirects users to a chat thread on WhatsApp, a widely used application. Keobot is exclusive to Keolis.

How do you guarantee reliable passenger information during public transit disruptions? This was the issue facing the Keolis Nederland network. Public transit issues are a source of dissatisfaction and stress for passengers and employees alike. The network thus wanted to set up an easy-to-use tool that would provide passengers with fast, reliable information. The adopted solution uses generative artificial intelligence. Based on observed events, the tool generates and adapts passenger information messages, which are then distributed via the network's communication channels: social networks, website, apps, on-board screens, etc. Thanks to this AI, information processing time has been cut from around 20 minutes to less than five minutes, with low investment and operating costs. The network's employees have already received training for this new tool.

The Keolis subsidiary Besançon Mobilités, Grand Besançon Métropole and local associations for the visually impaired have deployed the Ezymob mobile application on the metropolitan network (tramway, bus, coach). This digital assistant uses AI to provide real-time guidance, via audio and/or large-print visual indications. Three features are available: detection of access doors and empty seats, audio guidance at difficult transit hubs and a countdown of stops to the destination, with the aim of making mobility a little more accessible to all.

Anticipating peak traffic in real time thanks to artificial intelligence 

The Department of Transport and Planning in the State of Victoria, Australia, wanted to find an innovative approach to anticipating ridership on Melbourne's streetcar network and on public transit in general. To achieve this goal, Keolis Downer and the University of Melbourne joined forces to develop IMove. This tool uses artificial intelligence to analyze external data (weather, census, calendars, etc.) and data collected by ticket validation or counting systems installed on board certain vehicles. The algorithmic models they have developed make it possible to estimate tramway capacity in real time and more accurately forecast ridership. This innovation prevents the network from having to equip all trams with costly counting systems and makes up for the imperfect system of validation data, which does not take into account fraud and free travel in certain zones. In addition, by estimating ridership, we can adjust supply to meet demand and better coordinate the various modes of transport.

The optimal distribution of passengers in tramway or metro trains contributes to the quality of the passenger experience. It ensures greater on-board comfort and smoother boarding and exiting. In Lyon, Keolis and SYTRAL Mobilités – in partnership with the Affluences company – have experimented with video-analyzed passenger counting in each train, transmitting the information to the next station, where signals (green, orange, red) indicate the capacity of each approaching metro train. Passengers can then proceed to the doors of the least crowded cars before the train arrives.

Artificial intelligence to reduce wait times

How does Keolis ensure a waiting time of under 12 minutes for passengers on the shuttle bus to Mont-Saint-Michel? To do this, Keolis has developed a tool to anticipate peak traffic as accurately as possible. The shuttle buses transporting passengers from the coastal parking lots to Mont Saint-Michel now operate more efficiently, thanks to the integration of artificial intelligence algorithms. To implement this tool, Keolis' Northwest Regional Division has teamed up with Anjou-based start-up Cross Data. Powered by a wide range of data—including weather forecasts, calendars, parking barrier opening schedules, and Mont-Saint-Michel promotion campaigns—the tool forecasts passenger flows, adjusts transport provision accordingly, and provides real-time information on waiting times. This innovation improves the flow of user journeys, supported by universal signage featuring pictograms and convenient online booking tools. Forecasts can also be shared with other local stakeholders such as restaurant owners, whose business depends on visitor numbers. Ultimately, this approach makes the travel experience smoother and more inclusive, living up to the standards of this iconic World Heritage site.

Artificial intelligence in the service of safety

Keolis Amey Docklands has implemented an intelligent video surveillance solution to secure the tracks of London's Docklands Light Railway (DLR). The solution, called “CCTV AI Trial”, uses security cameras coupled with artificial intelligence to detect and identify objects on the tracks without mistaking them for train movements. An entirely new approach in the UK, the system improves incident prevention along railroad lines.

In Lyon and Manchester, Keolis is using artificial intelligence to enhance preventive maintenance on the overhead contact lines (OCLs) that supply the streetcars. Combining cameras, sensors and AI, this innovation monitors the condition of the equipment in real time and records every shock that occurs on the OCL. AI then classifies the data and indicates the main points on the network that require further inspection, before any serious damage occurs. In the context of operational excellence, this innovation reduces costly interventions and improves tramway safety and availability – all arguments in favor of more proactive maintenance.

Next steps: continue to explore the possibilities of AI and generative artificial intelligence 

Keolis is conducting research into the reconstitution of “origin-destination” journeys in collaboration with IVADO (Institut de Valorisation de la Donnée), led by the University of Montreal, and involving four Canadian university partners (Polytechnique Montreal, HEC Montreal, Laval University and McGill University). Thanks to artificial intelligence, Keolis is able to estimate passengers' exit points, since public transport generally only verifies boarding. By analyzing data from daily journeys, this initiative aims to refine our understanding of travel, while taking route variations into account.

Like many new technologies, the use of AI and generative artificial intelligence raises ethical and security issues, such as the protection of personal data, intellectual property and confidentiality. Keolis is now counting on artificial intelligence to transform mobility in France and abroad. These first practical applications will enable us to assess the benefits of this new technology and capitalize on feedback.