AI in transportation: moving faster towards the future

AI in transportation: moving faster towards the future

August 18, 2021


AI in transportation: moving faster towards the future

AI Researcher

As described by Homer in the Odyssey, it took Ulysses about ten years to return to his beloved Ithaca after the Trojan War, due to the wrath of Poseidon messing up with his travel plans. As men and women of the 21st century, we should accept this narrative involving angry gods and sea monsters with a pinch of salt, as it is more likely that Ulysses was simply better at designing wooden horses than at piloting a boat.

But this is not the point. Rather, have you ever wondered how things would have turned out if Ulysses had been able to count on modern transportation technologies? Could we assume that his journey would be much quicker and with fewer unexpected events? We will never know. What we do know for sure, however, is that humanity has made great strides in its ability to move using autonomous driving systems, GPS, and other navigation tools that have taken the place of the romantic but far less efficient stargazing.

In recent years, the rise of artificial intelligence has given impetus to this evolution and to the global smart transportation market which, according to MarketsandMarkets’ estimates, was valued at $94.5 billion in 2020 and may reach $156.5 billion by 2025 at a GAGR of 10.6%.

Global smart transportation market forecast, 2020-2025

Such excellent performance is also likely to be reflected in a new momentum for the global digital logistics market, which may greatly benefit from artificial intelligence solutions, rising from $17.4 billion in 2020 to $46.5 billion by 2025 at a CAGR of 21.7%.

Global digital logistics markets forecast, 2020-2025

Since we've just mentioned autonomous driving, let's take a brief overview of the major embodiments of artificial intelligence in transportation, starting right from this fascinating technology.

1. Do self-driving cars need a license?

Autonomous driving is probably one of the most exciting and, at the same time, most controversial aspects of implementing AI in transportation, leading public opinion to split, especially in the past, between those who saw it as a miracle and those who considered it as dangerous as witchcraft.

In this regard, Deloitte's 2020 Global Automotive Consumer Study reported that around half of the American consumers surveyed were skeptical about AI-powered self-driving vehicles' safety, while the Chinese appeared to be more confident in this technology.

Consumer opinion on autonomous vehicles safety by country

While it's wise to address such concerns as constructively as possible, the irony lies in the fact that many consumers do not realize how much the adoption of similar AI-related technologies is already rooted and widely accepted in transportation sectors such as civil aviation. According to the New York Times, for example, the actual human control during a typical Boeing flight lasts about 7 minutes and is almost exclusively limited to take-off and landing, while the autopilot system takes care of everything else.

Soon what is absolutely common in aviation may be the new standard in all other sub-branches of transportation, thanks to the latest advances in artificial intelligence, such as in deep learning and computer vision. By leveraging these cutting-edge technologies, machines are acquiring capabilities that were previously a privilege of humans, such as the ability to learn from experience, make decisions on their own, recognize objects and people around them, and even more.

Similar solutions have already been implemented in a variety of AI use cases involving both commercial and personal transportation. Let’s have a look at some relevant examples of this growing trend.

Passenger motor vehicles

Many car manufacturers have long started to implement semi-autonomous driving features in their vehicles, such as advanced driver-assistance systems (ADAS), to help perform parking procedures or ensure control of the vehicle in bad weather conditions. However, the possibilities unlocked by artificial intelligence in transportation go far beyond that.

In Tokyo, for example, self-driving taxis produced by Japanese corporation ZMP have already been a reality since 2018, although a human operator inside the vehicle is still required for safety reasons. In the United States, on the other hand, California-based development company Waimo opened its robotaxi ride-hailing service to customers in 2019.

Commercial motor vehicles

Moving on to road freight vehicles, AI proved to be a valuable ally as well, but its potential is far from being fully harnessed. Based on McKinsey's estimates, the large-scale deployment of autonomous trucks (ATs) in a country like the United States, where 65% of freight is transported by road, could lead to the truck rental industry cutting its operating costs by 45% and saving up to $125 billion.

Such achievements would be possible, for example, by using AI-based autonomous systems for truck platooning, i.e. the coordinated movement of multiple trucks at close range, capable of advancing or braking almost simultaneously. This caravan of vehicles may be led by a human driver controlling the first truck, while the following trucks wouldn't require active driving but mere supervision in case of unexpected events.

Example of truck platooning

Regarding self-driving trucks, it is also worth mentioning that such AI-powered vehicles can be leveraged not only for standard road transportation but in much more "hardcore" contexts too. Australian mining corporation Rio Tinto, for example, deployed autonomous trucks in its sites in Pilbara to transport minerals and waste, achieving a 15% reduction in load and haul unit costs. Each vehicle was able to operate 700 hours longer than their standard counterparts and no miners were injured during the operations, proving that such solutions are both cheaper and safer.

Intelligent trains

Artificial intelligence in transportation is also making its way into the rail sector, incarnating in the so-called ATO (automatic train operation). In this regard, the 2019 Artificial Intelligence in Transport paper drawn up for the European Commission reported that, in 2018, the inhabitants of 41 cities in 19 countries around the world could already count on a total of around 1,000 km of driverless metro lines, and their extent seems set to grow to over 2,300 km by 2025.

Cargo ships

We've already talked about land transportation, but what about the sea? Well, AI-powered autonomous ships could be another trump card for logistics companies to improve efficiency and reduce emissions. The American tech corporation Sea Machines Robotics and the world’s major shipping line Maersk, for example, partnered to equip container vessels with video object recognition and LIDAR (light detection and ranging) technologies. Such tools ensured autonomous navigation and route optimization and helped to cut operational costs by 40%.

2. Traffic police or… AI?

Having mentioned AI-powered route optimization for cargo ships, it might be worth extending this topic to all those advanced tools that help us rationalize our movements, harmonize traffic, and coordinate fleets of vehicles, ships, and planes.

Traffic management systems might be seen as something less glamorous and sci-fi than other high-tech transportation solutions such as autonomous vehicles (they’ve made several films and TV series with driverless cars but not so many with intelligent traffic lights as protagonists, as far as I remember). However, smart fleet and traffic management systems still represent one of the most useful manifestations of artificial intelligence in transportation and logistics, as they greatly improve data-driven decision-making.

Their operation is based on a blend of GPS, sensors, computer vision-powered cameras, and other interconnected IoT devices deployed to gather data regarding weather, traffic, blockages, or accidents. These tools are then combined with AI-based analytical systems to process such information, identify recurring traffic patterns via machine learning algorithms, and turn data into valuable route recommendations or forecasts on potential congestions.

Actually, we all got a taste of these transportation-related technologies by using Google Maps to avoid queues and get tips on the shortest route while driving in the city. But how can we leverage it on a bigger scale? Here are some interesting case studies.

On the road

An AI-based monitoring system developed by Siemens Mobility has been deployed in Bengaluru, India, to collect data with smart cameras, adjust traffic lights in real time, and thus facilitate traffic flow. A similar solution, called SurTrac and created by Rapid Flow Technologies, has also been adopted in Pittsburgh, US, leading to a 25% reduction in travel times and a 20% cut in emissions. A perfect example of how artificial intelligence may represent a key factor in achieving sustainable transportation.

Example of a real-time traffic optimization system

In the skies

To improve its air traffic control and deal with the infamous London weather, Heathrow Airport has implemented Aimee, an artificial intelligence system fed with data collected by multiple high-definition cameras and capable of assisting controllers in supervising arrivals and departures. Once deployed at full capacity, this tool should be able to enhance the airport's landing capacity by 20%.

Across the seas

Returning to the realm of Poseidon, AI in transportation is already widely used not only for the aforementioned autonomous navigation but also for the optimization of port procedures and traffic. The Port of Rotterdam, for example, has adopted an AI-based system to estimate ship arrival and departure times, leading to a 20% reduction in waiting times and therefore ensuring a significant cut in operating and fuel costs.

3. My mechanic is an algorithm

Self-driving vehicles and AI-based traffic management systems are amazing tools... until they break down. In this regard, a rapidly spreading approach in transportation and other industries, known as predictive maintenance, is based on the idea of forecasting potential failures in advance before they actually occur.

Nowadays, the efficiency of this modus operandi has been greatly enhanced by advances in artificial intelligence and, in particular, thanks to the aforementioned pattern recognition and prediction capabilities of machine learning algorithms. Indeed, a machine learning system can be trained to understand the normal operation of vehicles and infrastructures by "feeding" it with data relating to their standard performance and collected via sensors.

Once the system has learned to recognize the ideal operating patterns of the mechanical and electronic components powering a vehicle or an entire station, it will also be able to detect any alteration in their typical behavior that might be a sign of impending breakdown.

By implementing anomaly detection via machine learning for condition monitoring, transportation operators can significantly improve vehicle reliability, reduce maintenance costs, speed up repair procedures, and cut fleet reserves normally deployed to avoid service interruptions in case of failures.

A striking example of this approach comes from France's national railway company SNCF, which implemented AI-driven predictive analytics to monitor the status of pantographs at risk of wear and ensure proper electrical power supply to its trains. According to the company, predictive maintenance will also lead to a 30% reduction in accidents related to train switches.

Let the machines guide us

While the transportation industry does not lead the ranking of artificial intelligence implementation, it still represents one of the sectors benefiting from the highest adoption rate of AI-related technologies, as reported by McKinsey's 2019 Global AI Survey.

AI adoption by industry

This should come as no surprise, as AI in transportation proved to be capable of making us move faster, safer, and cleaner than ever before. However, deployment efforts should always be carried out with a bunch of essential requirements and controversial factors in mind:

  • Setting up a long-term strategy that covers investments necessary to integrate AI into the pre-existing technology stack and business processes.
  • Defining specific use cases involving critical areas of a transportation business that could greatly benefit from AI-based enhancements.
  • Starting small and scaling, in order to achieve fast results in the short-term and overcome the fears and resistance of employees and shareholders.
  • Investing in training and reskilling to reallocate the workforce which may be displaced by the advent of machines.
  • Increasing cybersecurity defenses, as relying on technology and data also increases the vulnerability of transportation networks to potential cyberattacks.
  • Addressing some ethical and practical implications of autonomous driving and ensure that AI-driven transports can safely interact with their passengers, standard vehicles nearby, and the surrounding environment. This also relates to making decisions based on a not-so-well-defined concept of lesser evil in emergency situations.

Besides these pragmatic aspects, we could add the skepticism among the most romantic travelers, who might be worried about missing out on the pleasure of actually driving or the more adventurous aspects of their journeys. But if it’s true, as Lord of the Rings author J. R. R. Tolkien said, that "not all those who wander are lost," it's also undeniable that sometimes you don't want to wander but just go from A to B in the fastest and easiest possible way. And artificial intelligence may be the best companion to help you find the right path.