by Andrea Bergonzi, Data Scientist of Dataskills

Until a few years ago, when we talked about the use of Artificial Intelligence in the Transport sector, the main application that was mentioned was predictive maintenance. The latter, which obviously continues to be extraordinarily crucial, allows maintenance activities of a train, an aircraft or a car to proceed following a prediction of the so-called “breakdown time” (i.e. the remaining operating time) of the critical components of the vehicle. Predictive maintenance is an evolution – but also, intrinsically, a conceptual contrast – of traditional reactive maintenance, which involves repairing a vehicle only after a fault has occurred, and preventive maintenance, based on checkups at constant intervals and regardless of the state of wear of the vehicle.
However, with the dizzying evolution of Artificial Intelligence and its progressive penetration into practically every sector and context of our life, its application potential in the Transport industry has also expanded considerably, simplifying activities that generally require human intervention.
These include:

• Safety: with the presence of sensors on vehicles, which increases the possibility of preventing accidents thanks to the processing of driving data and the evaluation of the driver’s state of attention.
• Performance: AI allows you to save time and reduce polluting emissions by monitoring traffic and indicating to users the best routes to follow. The creation of predictive models that are “shaped” on real traffic flows also proves to be very effective for optimizing consumption.
• Network infrastructures: the installation of specific sensors allows identifying potential improvements on roads in order to reduce the risk of accidents and optimize traffic. Sensors can also be used to monitor tunnels, bridges and even the road surface, so as to resolve problems before they become critical.
  

And of course there is much more, as demonstrated for example by Truck Platooning. This project intends to revolutionize road transport through a technology that allows the travel of a convoy of two or more heavy vehicles (specifically articulated vehicles) coordinated via wireless. The goal is to obtain, by 2025, a “platoon” of over three trucks driven by a single driver, simultaneously reducing greenhouse gas emissions.

In the near future, AI in the transport sector will have to respond in particular to users’ needs for efficiency, safety and convenience. And if, in this sense, innovations are already starting to be seen with technologies such as autonomous driving, integrated assistance and payment, IoT and the connection between vehicles, the medium-long term objective remains the creation of a complete ecosystem of intermodal, sustainable and service-oriented mobility, which supports safety, modernizes infrastructures and makes them more sustainable, maintenance efficient and offers concrete assistance to people (be they operators, drivers or pedestrians).
These are ambitious goals and, not surprisingly, the AI market in the Transport sector is already very large: according to Precedence Research, it reached 2.3 billion dollars in 2021 and could reach around 15 thousand by 2030, with an annual growth rate of almost 23%.
Let’s see below which are the AI technologies and applications that could be destined to have the greatest impact on the transport of the future.

Autonomous vehicles: reshaping the way we move
After a “quiet” start, caused by slower technological development than initially expected, autonomous vehicles are now a reality (although not yet very widespread) and could change the way we move and use roads and systems. transport.
And it’s not just Tesla that is focusing on this application: Alphabet (Google), Waymu and Baidu have also already implemented such systems and continue to invest in innovation. Elon Musk’s giant, for example, has been offering driving assistance functionality (called autopilot) for some time, while level 5 AV technology (i.e. totally autonomous driving), considered the true competitive lever of the brand, for the moment is not yet producing the desired results. Meanwhile, Baidu offers driverless taxis in China and plans to build the world’s largest self-driving ride-hailing area.
For all these companies, autonomous driving applications are based on Artificial Intelligence tools such as artificial vision (or Computer Vision): in this case the AI cameras are called upon to interpret the data in real time in order to allow the vehicle to avoid pedestrians and obstacles, maintain a cruising speed in compliance with the law and reach your destination easily and safely.

Intelligent traffic management: to improve traffic
The use of AI in transport could also prove to be a keystone in intelligent traffic management, through the use of smart traffic lights that switch from red to green in the absence of traffic in a given direction.
In this context, the use of Artificial Intelligence sensors and “traffic models” could also determine the best ways of regulating traffic during peak times, for example by temporarily transforming two-way streets into one-way streets to favor traffic flow. vehicle conveyance.
At the same time, Artificial Intelligence algorithms could also prove useful in predicting moments of particular congestion, improving traffic flow, generating concrete time savings, reducing air pollution and accelerating the transport of goods.

Road accident detection: for faster and more efficient management
Real-time detection of road accidents is another of the applications considered most interesting for Artificial Intelligence in the field of Transport. The objective, as is easy to imagine, is to minimize inconvenience to motorists and at the same time increase their safety.
If road monitoring has long been the task of video surveillance, AI could soon take its place in the real-time observation of road networks and intersections, in order to ensure quicker management of any accidents. The main advantage of the application of this technology is linked to overcoming “human” limits in the field of surveillance: it is in fact impossible for one person to monitor multiple cameras simultaneously and with the same efficiency, with the result that incidents are not always detected and managed at crucial moments.
In contrast, the computer vision system has virtually no limits in looking for rear-end collisions, queues and unusual traffic conditions as it monitors all cameras. Furthermore, a further advantage of this application is the increase in the prediction of possible critical issues in particular key areas.

Traffic and parking regulation: for more punctual application of the law
In the transport sector, Artificial Intelligence is destined to have a positive impact also in the application of rules relating to roads and parking. By relying on IoT sensors and cameras for data collection, it will be possible to detect the occupancy status of parking areas, immediately informing motorists of any “full” areas and suggesting alternative solutions for parking their vehicle – more quickly and reducing road congestion.
At the same time, AI systems will be able to inform the authorities about vehicles that have violated the highway code. As? By detecting transit speed, entry into restricted traffic areas or parking in prohibited areas. The automatic recognition of license plates and the detection of the model and color of the vehicle will allow quicker identification and, consequently, more precise charging of the violation committed.

Artificial Intelligence and Transport: a combination destined to evolve
Ultimately, the application potential of Artificial Intelligence in the Transportation sector is almost limitless, and current technologies seem to have only scratched the surface of what is possible.
In the coming years – and almost certainly decades – these technologies will be implemented more and more and will progressively take on tasks and activities that were previously reserved only for human beings. Should the researchers’ hypotheses prove to be well-founded, these important innovations will completely eliminate human errors, automate and simplify numerous processes and accurately predict future needs. The impact on the transport sector will be enormous, thanks to the ability of these technologies to totally revolutionize the way we move from point A to point B.

And while it is true that this transition will take time, the change will be inevitable: it should not be forgotten that Artificial Intelligence is in fact an integral part of future smart cities, i.e. new urban ecosystems that emphasize the use of digital technology and shared knowledge for the benefit of public safety, health, mobility and collective productivity.