A real-time decision support system for reducing CO2
and black carbon emissions

Testing for black carbon is a common way to determine levels of traffic exhaust emissions, which is linked to higher risks of cancer, heart disease and lung disease and is also known to exacerbate allergies and asthma. With the trend of increased urbanisation, traffic congestion — a major contributor of CO2 and black carbon — is also increasing as the rate of urbanisation outgrows the rate of infrastructure development.

This trend brings major challenges not only to the environment but to healthcare and social environments as public infrastructures are currently stretched to their limits.

A recent study estimated that the monetised value of particulate matter PM2.5-related mortality attributable to congestion in the year 2000 in 83 US cities was approximately $30 billion.

Another study showed that every 10 micrograms / m3 fine particulate matter elevated the risk of developing cardiopulmonary disease (6% increase) and lung cancer (8% increase).

The equivalent of air pollution during the Victorian era was the waste volume and sewage problem. As it was then, a revolution in sanitation introduced by Edwin Chadwick improved the health and well-being of the people. Similarly, a disruptive solution and new ways of thinking are required to meet the trend of increased urbanisation.

Carbotraf EU projectIBM Research – Ireland participated in an European Union-funded research project called CARBOTRAF to design a real-time predictive analytics monitoring system for traffic operators to reduce air pollutants caused by traffic congestion. The city of Graz, Austria, contributed data to this project in real-time to assist in the decision-making process of the cities' traffic operators.

To ensure that our cities continue their development as economic hubs and retain the ability to connect urban as well as suburban life, a paradigm shift in tackling our infrastructural problems is waiting to happen. As Edward Glaeser argues in a recent Science article, “Clean water [in big American cities] was solved with an engineering solution, it has become clear that we cannot build our way out of traffic congestion”.

Historically, public infrastructure has emerged without a guiding global design, but rather is based on an on-demand structure and near-future outlook. In accordance with continuous physical improvement on a city's infrastructure, technological advances promise to circumvent some of these physical limitations.

Traditional reactive control methods for public infrastructures are woefully inadequate as current monitoring of traffic control decisions in a city are often too late to mitigate traffic jams. This of course has a detrimental impact on the main benefits of city living: the ability to connect easily with each other for the exchange of ideas or goods.

Resigning oneself to the idea that megacities are chaotic and disordered systems is simply untenable. The proverbial butterfly in the Amazon causing a tornado in the United States prompts us to surrender to the notion that minute perturbations in the state of the world can lead to massive consequences.

Research using real-time predictive analytics is the sophisticated, practical and affordable game changer in traffic monitoring as it combines real-time data and information analysis tools to develop traffic forecasts and models that are helping to reduce congestion, CO2 levels, and black carbon to produce a better quality of life in cities.

Contact

Dominik Dahlem

Contributors

Dominik Dahlem Rahul Nair Tigran  Tchrakian John Sheehan Sergiy Zhuk Eric Bouillet
Olivier  Verscheure Susara van den Heever Alessandra Pascale
Challenge

Provide real-time decision support for traffic operators to improve traffic management as well as to reduce CO2 and black carbon emissions.

Solution
Benefits

Use case: Deployment in Graz, Austria

In order to foster a stronger integration of intelligent traffic management while at the same time meeting important environmental objectives, we developed a real-time system that accounts for both environmental and traffic considerations.

The CARBOTRAF solution assists traffic operators to reduce CO2 and black carbon emissions caused by urban traffic through an adaptive traffic management system. The system ingests real-time data on traffic and meteorological conditions and informs traffic operators about potential environmental as well as traffic impacts of traffic plans.

For this it relies on statistical models that are built offline in order to score and rank available traffic plans according to key performance indicators. This allows traffic operators to actuate city infrastructures, thus reducing emissions while meeting traffic-related service level agreements.

The system is built on IBM Infosphere Streams, a scalable low-latency middleware platform, and relies on a family of statistical models for CO2 and black carbon emission, travel time, and air-quality predictions at defined hot-spots. The predictions from these models assist traffic operators to select traffic plans that result in the most significant reductions of emissions while at the same time maintaining traffic-related service levels.

Reducing emissions

Deployment

City of Graz (right): Balance the load between two main south-bound arteries into the city.

Reducing emissions

We demonstrate the feasibility of linking real-time traffic control decisions to environmental impacts. We deployed our system in the city of Graz, Austria, and showed that traffic and environmental indicators can be utilised together to inform traffic operators reliably about the best-performing traffic plans.

Interestingly, we were able to show that novel plans incorporating a traffic information system for inbound travellers have the potential to improve upon the currently deployed on-peak ITS plan.

This potential improvement was validated and confirmed in our case study in Graz, where a reduction of 1% was observed, which amounted to 10 tons of CO2 and 1 kg of black carbon over a 3-month evaluation period for morning peak traffic (between 6-9 a.m.).

The mathematical tools comprising the decision-support system are applicable to domains other than transport where real-time decisions are needed to improve operational characteristics.


Application areas

Smarter water
Smarter energy
Smarter healthcare
Smarter transportation
Smarter cities