Road vehicles are a significant source of pollution in Canada, accounting for about 145.1 megatonnes of carbon dioxide equivalent in 2016, or about 21 percent of total greenhouse gas emissions.
As a result, municipalities are beginning to consider the environmental costs of vehicle emissions as part of their traffic management practices. The Region of Waterloo in Ontario, for example, looks at fuel consumption and emissions when conducting intersection control studies.
There is currently no way to efficiently measure emissions as a direct result of vehicular traffic, because gases released from vehicles immediately begin mixing with other pollutants in the atmosphere. A portable emissions measurement system on every vehicle would be the best solution, but is impractical at the moment. A more realistic solution is to estimate vehicle emissions by modelling data collected from traffic management systems, along with other sources.
Miovision, the University of Waterloo and ODX came together to address this problem and research a solution.
Miovision, a smart city company in Waterloo Region, creates traffic-management technology, and has been successfully implementing its systems around the world for more than a decade. The opportunity to measure emissions impact on traffic management would further the company’s leadership position. A team led by Dr. Liping Fu from the University of Waterloo’s Civil and Environmental Engineering department was brought on board to conduct the research in estimating vehicle emissions.
A stretch of Hespeler Road in Cambridge, Ont. was selected as the test site, and the team identified several sources of data they would need to feed into the emission models. Miovision provided traffic count data, hi-res signal timing, and travel time information from existing sensors in the area. For intersections where traffic counts were not available, the Region of Waterloo provided estimates, along with information on signal plans for intersections along the route. And open data sources such as weather and vehicle age distributions were used as inputs into the models. Using open data allowed the team to incorporate important information free of any restrictions.
The data was fed into a first model, Vissim, that simulated traffic activities. The model was refined and validated to site measurements to ensure data was realistic. The output from this was then fed into a second model, MOVES, along with weather, vehicle data and fuel economy data. This information would allow the emissions simulations to be run. Using
the simulation results, simplified models were developed to estimate emissions without having
to run simulations every time.
The project was able to generate a new, efficient solution for tracking emissions from road vehicles that can be used by Miovision, enabling the company to enhance its offering to existing and future customers.
Miovision’s Sajad Shiravi summed up the benefits of this project: “The models developed in this project enable us to estimate emissions from vehicle trajectory data extracted through our AI-enabled computer vision technology at the intersections. This can now be possible without the need for an extra piece of air quality measurement hardware at the intersection and allows us to further enhance our toolkit of traffic performance measures.”
As an additional benefit, the data and models were opened up to allow others to build on this research. Developing models can be a time-consuming task for research projects. By making these openly available, it lowers the barriers for others to take on new projects. The data can also be used by others to explore new ideas or methods for tackling the problem of greenhouse gas emissions from vehicles.
The Miovision/University of Waterloo project is an example of how data can be utilized to address significant challenges. The data and models generated through this research can help municipalities better understand the impact of traffic signal timing on greenhouse gas emissions from vehicles. And with the data openly available, others have the opportunity to further explore methods for reducing vehicle emissions. While this may not eliminate the problem, it provides one more step to reducing the environmental impact of vehicular traffic.
Links to the datasets, along with a Tableau visualization created for the project, are included below.
Tableau Visualization
Datasets and Models on GitHub
Full Data Files
