AI in transportation

Using machine learning to improve mobility

In today's rapidly evolving technological landscape, one field that has captured widespread
attention is artificial intelligence (AI). With its ability to replicate human intelligence and
perform complex tasks, AI holds immense potential for transforming various aspects of our
lives. In the following, we concentrate on the essence of AI in transportation, its impacts and
its benefits.

What is artificial intelligence – and is it here to stay?

Artificial Intelligence (AI) is one of the most exciting and rapidly developing fields of
computer science. It involves creating intelligent technology that can perform tasks that
typically require human intelligence, such as learning, problem-solving, decision-making,
and natural language understanding.

While there are concerns about the impact of AI on jobs and society, many experts believe
that the benefits of AI will far outweigh the risks. By automating routine tasks, AI can free up
human workers to focus on more creative and meaningful work. Moreover, AI can help us
make more informed decisions and better understand what surrounds us.

The applications of AI are diverse and far-reaching. Let’s explore its effects on
transportation.

Artificial Intelligence in transportation

Artificial Intelligence (AI) is revolutionizing the transportation industry, leading to
significant improvements in efficiency, safety, and convenience. To achieve this, machine
learning is becoming more common in many sectors of mobility.

In the past few years, advancements in machine learning have sped up the adoptionof AI in
many areas, including transportation and mobility.

So how can mobility planners and operators - and above all the public - benefit from AI?
Here are a few examples.

AI and autonomous vehicles

One of the key transportation areas where AI is having great impact is in autonomous
vehicles. Self-driving cars have the potential to reduce accidents caused by human error and
improve overall traffic flow. Many major car manufacturers and technology companies are
currently developing autonomous vehicles, with some already testing them on public roads.
In many cases, developers train AI control algorithms to reproduce the behavior of
experienced drivers, navigating through surrounding traffic. PTV Group, for example,
collaborates with AI developers in projects like CoExist to ensure that autonomous vehicles’
behavior is simulated correctly in PTV's traffic simulation products. AI driven gadgets
(ADAS) optimize the way we drive and reduce the possibility of human errors.

Autonomous self-driving cars would have to be widely adopted for smart roads to unlock
many of their benefits. At the moment, the public is a bit skeptical about handing over the
steering wheel to artificial intelligence. The gridlock in regulations and legislation reflects
public distrust.

AI in traffic management systems

AI is also being used in traffic management systems to optimize traffic flow and reduce
congestion. By analyzing real-time traffic data, AI algorithms can adjust traffic signals and
reroute vehicles to less congested roads, reducing travel time and fuel consumption.

Intelligent traffic management is already being implemented successfully. Cities

like Taichung, Vienna, York, or Rome already rely on PTV’s real-time solution which
combines machine learning techniques with dynamic traffic modeling.

This software, PTV Optima, helps operators to make reliable traffic forecasts up to 60
minutes in advance and identify the best scenarios to manage congestion, road closures and
construction sites.

Smart traffic management enables operators to quickly react to changing conditions. This can
prevent or mitigate traffic jams and congested roads, even before they happen.

AI also helps to improve the monitoring of traffic volumes in many locations across the
network. AI helps traffic controllers to spot abnormal traffic conditions and bases its forecasts
on similar historic traffic situations.

AI can also be used for adaptive signal control. For example, Taipei uses PTV Balance, a
software that continually optimizes traffic lights, to improve traffic flow and to minimize
emissions and vehicle delays.

Eco-friendly mobility is often powered by AI and smart technology. The aim is to obtain and
process data and improve how transportation systems work and interact with each other. If
these systems were implemented more often, we could experience their immense benefits.

AI in public transport
Just like private transport, real-time management of public transport can benefit from AI-
assisted optimization. Big data streams from ticketing systems and automated passenger
counting equipment contribute to a better understanding of passenger flows through the
network. Based on improved situational awareness, traffic controllers can take action when
demand patterns deviate from average conditions as well as when delays or infrastructure
failures affect operations.
Algorithms can assist both traffic controllers and passengers with suggestions how to best
cope with the situation. Traffic controllers will be able to decide on the best actions for
recovering normal service, e.g. by short-turning trains or buses, substituting buses for trains,
or selecting which connections to cut and which to maintain, in order to minimize the impact
on passengers. Passengers receive notifications describing the best alternative route to their
destination which minimizes delay on arrival.

Implementing all of this as part of PTV Optima has important advantages: not only are the
decisions based on one overall traffic state for public and private transport – essential for
practical bus substitution. Journey planner recommendations to passengers are also always
consistent with the dispatching actions of traffic controllers, stranding fewer travellers on
their journeys.

Currently PTV is working on the implementation of this objective within the EU-funded
research project UPPER, coordinated by UITP.

AI in shared mobility
For mobility-on-demand services, AI can optimize the deployment of shared vehicle fleets
and improve the user experience of passengers.

By analyzing data on passenger demand and traffic conditions, AI algorithms can predict
passenger demand up to one hour ahead. Idle vehicles are then sent to future demand
hotspots, just in time to pick up passengers. This reduces waiting times and detours.

In this case, AI beats conventional time series analysis: neural networks not only look at the
temporal evolution of demand, but they can also discover spatial patterns: often the demand
in one neighborhood is correlated with that in adjacent parts of the city.

AI in action: The story of #transmove

Germany’s second largest city Hamburg strives to overcome congestions, reduce emissions
and improve mobility. Therefore, a project funded by the German government and the City of
Hamburg, leads the way: #transmove. The aim is to enable smart and sustainable mobility
forecasts and to issue recommendations for better traffic management.

#transmove enables traffic and city planners as well as ordinary citizens, to reliable live long-
and short-term mobility forecasts – directly to their computers or mobile devices. This
contributes to a steady traffic flow and improved mobility, which goes hand in hand with
reduction of harmful pollutants from transportation.

In #transmove, the mobility forecasts are generated through an innovative blending in of

dynamic agent-based simulations (modeling individual behavior of mobility users) with static
forecasting approach based on the Hamburg transportation model. The results and proposed
actions are made available to all stakeholders, meeting their differing needs and
expectations.
In addition to an integrated machine learning algorithm, a simulation of individual mobility
behavior of users is included in the forecast calculation as part of the accompanying scientific
research on agent-based modelling.

Based on the software, traffic coordinators can assess the traffic impact of planned measures
on the mobility flow in Hamburg. Thus, ideal times (e.g. in times with lower traffic volumes,
parallel implementation of construction measures) can be forecasted for an efficient planning
of infrastructure work. This saves the traffic coordinators time and costs, because all data
relevant for planning is available in one software . The advantage is also in quality: The
mobility forecasts based on AI methods are very precise and decisions for traffic coordination
measures can be made based on a very realistic forecast.

In short: planners and decision-makers get better quality of information and forecasting;
citizens get better quality of life.

AI in transport logistics
Another area where AI is making a difference is in logistics and supply chain management.
By analyzing data on shipping routes, traffic patterns, and weather conditions, AI algorithms
can optimize delivery routes, reducing fuel consumption and emissions.

Will human dispatchers in logistics companies soon be replaced by AI-powered transport

planning software? The answer is clearly No. Routing and scheduling software uses
algorithms to calculate routes. But not all variables, restrictions and conditions can be
mapped by algorithms. Software also does not consider exceptions and spontaneous
deviations. This is where dispatchers come into play with their knowledge and experience.

Future of AI in transportation
Overall, AI is transforming the transportation industry, making it more efficient, safe, and
convenient for everyone. While there are still challenges to be overcome, such as the need for
more data and the development of robust regulations, the benefits of AI in transportation are
clear and will continue to grow in the future.

But there is huge room for growth. Today’s availability of big data is so far largely untapped,
so there is massive potential. Not only in terms of machine learning and AI, but also in data
analytics and visualization.

At PTV Group, we are working on a series of new dashboards to be released soon, that will
deliver easily accessible mobility insights to cities. In 2022, PTV Group released the first
dashboard, PTV Access, to visualize the accessibility and mobility scores in German cities-
French and cities around the world coming soon. Such tools empower city stakeholders to
shape sustainable and inclusive urban environments focused on the citizen’s needs.

In conclusion, AI in transportation can help us to create a system that is safer, more efficient,
and more sustainable. With continued research and development, we can look forward to a
future where transportation is more accessible, convenient, and sustainable