Current State of Affairs
The basic premise of this project is to use traffic modeling to investigate traffic routing in Ottawa, especially concerning the link between the Quebec route 5 and the Ontario 417 through the downtown area.
An important aspect of urban design that interests me greatly is the “reclaiming” of the city centre for pedestrians. This can be accomplished in a great number of ways, including the rerouting of traffic, building biking lanes, sidewalks, and public transit. It is something that is needed desperately in our North American cities, as the automobile based city design is simply not sustainable going forward, and we need to save and rejuvenate our city centres in order to keep them from becoming another Detroit, Cleveland, or Los Angeles. It is therefore important for us as citizens to look around us in our own cities, and observe with open eyes what things can be improved. It is important to be critical of the design of our infrastructure, taking steps to promote the redesign where necessary.
Due to the unique geography of Ottawa, There is much need for access between the Quebec highways to the north of the river, and the Ontario highways on the South side. The major Quebec highways leading to the Ottawa-Gatineau metro area are the 50 and the 5. The 50 begins (or ends) in the centre of Gatineau, and continues to a junction just north of Montreal. It is the main Quebec link between the Capital area and the Montreal metro area, and is also the main Highway that all Quebec towns feed into between the two large metro areas. The highways to the west of the Gatineau urban area are smaller than the 50, but also contribute to traffic entering into the city.
The Quebec route 5 is a shorter highway extending to the north of Gatineau, which connects it to the Quebec towns of wakefield and Chelsea, as well as the areas further north. This highway meets with the 50 in downtown Gatineau, where it crosses the Ottawa river into Ontario.
On the South side of the river, the main highways of Ontario are the 417 and 416. The 417 is the main East-West highway in Ontario, and is the Ontario section of the famous Trans-Canada highway. It crosses the city quite close to the centre, and continues, with a few branching segments towards the suburbs of Ottawa. The 416 (also part of the Trans-Canada) extends south between Ottawa and the cities of Kingston and, further on, Toronto.
The problem that Ottawa has lies in the connections between the two major highway systems. The major provincial highways are not connected with any high-flow highway system. This necessitates the movement of cars, and especially large trucks, through the small streets in the centre of the city to cross from one highway system to the other.
We can see that as cars and trucks cross from the Quebec side to the Ontario side, they search for, and take the shortest street path to the on-ramp of the 417 near the Lees neighborhood. This involves driving down King Edward avenue, followed by Rideau street through the entertainment district of Ottawa, then past the Rideau Centre mall, where there are many pedestrians, and finally along Nicholas street in between the University of Ottawa campus and the historic Rideau Canal.
Figure 1: the route taken between the major highway systems.
Figure 2: a closer view of the route as it passes through the city on the smaller city streets. University of Ottawa is shown in yellow, while the ideally pedestrian friendly Byward market area is shown in green.
Clearly, there is a better solution for the routing of traffic, as the current state of affairs is far from ideal.
Traffic Modeling
This traffic routing problem is an interesting situation, because there are many ways in which the route can be changed. For this example, we have 4 (and maybe more) realistic solutions to this traffic route. The first is to route the large traffic through St. Patricks street and across to take the Vanier Parkway to the Hurdman onramp. The second is building bridges to the east, to connect the highways across the river, a third might be to build a link between King Edward avenue and St Patricks through the Bordeleau park. And finally a fourth solution would be to build bridges to the west, and extend highway 50 further through Hull and across onto a connecting highway in the region of Westboro or Bayshore. These solutions are all interesting, and certainly worth looking into. But the most realistic option, and the most interesting to try and model is the St. Patricks street diversion onto the Vanier parkway.
This option allows investigation into some traffic modeling software, and delving into the intersection of GIS, urban planning, and traffic design. Before looking at the simulation, a quick map shows the route that I am proposing.
Figure 3: the new route, avoiding the central areas of Ottawa.
Importing
The first step in the traffic modeling is to find a good software that can build and simulate the local traffic situation. I chose to use InfraWorks 360, which is part of the Autodesk line. Partially this was because it is free to students, and partially because it uses the same functionality as autoCAD and other 3D modeling softwares, so there is good user involvement and resources to draw upon. Right off the bat, one of the interesting aspects of the software is the fact that it is cloud based, so the entire workflow is extremely minimalistic and uncluttered.
The workflow of the traffic modeling starts with selecting an AOI from an open source map, such as google maps, and importing it into the program.
Figure 5: the import selection
The program renders it and automatically adds roadways, terrain heights, and 3D buildings.
Figure 6 : the imported terrain and 3D buildings.
I turned off the buildings layer immediately, as it is irrelevant for this project, and focused on the terrain and roads layers. Unfortunately, the automated rendering of both terrain and roads is quite far off of reality. Terrain is also not relevant to the project, so i will leave that as being inaccurate. However, in the initial inspection of the roads layer, major discrepancies were apparent.
Data Cleaning
When working with traffic simulations, there are a few key layer attributes that are most important. Firstly the road type, with the sub-attributes of number of lanes, speed limit, and width. Secondly there is the intersection layer. This layer has the attributes which include signal type, flow restrictions (such as right of way, yield, etc), intersection timing, and demand. Both of these layers need to be setup with their attributes matching as closely as possible to real life. Unfortunately the automatic render from the import of the map file is not accurate, so the job of cleaning the data set is extremely important. This process makes up the bulk of the workflow for the simulation. I found this process actually more involved and demanding than any GIS data cleaning that I have done in the past, due to the setup of the program. Instead of attribute tables and numbers that one might usually be able to clean through SQL and other traditional methods, this is a much more visual and hands on process. It also highlights how complex our road network is in real life. Hundreds of lane changes, one ways, left turn forbiddens etc.. All of these had to be entered in. luckily for traffic simulation purposes, only the streets of the simulation AOI needed to be correct.
Firstly, I went through the map and deleted features that were clutter, or not relevant to the simulation such as walking paths. Then I began cleaning the streets that were parts of the simulation route. The best method I found to accomplish this stage was to split my computer screen with both Google Street-view and Google Maps open on one half, and the program running on the other half. I referenced the Google Maps image of each street segment, looking at intersection types, lanes, width, and road style, and used the edit geometry tool on the Infraworks program to create streets and intersections that matched real life. Firstly to create the street geometry, and secondly to apply the street rules (one way, speed limits, lanes, etc). I used Ottawa city data to find speed limits and traffic light locations. The positive side to this is that we can narrow our AOI when we run the simulation to only include the cleaned and referenced streets, which means that any local streets that did not connect to my study area can be left incorrectly rendered. This process is long, and extremely finicky. Even with my relatively small study area the process takes days, and it is not something that can be automated (to my knowledge).
Simulation Setup
In looking through the likely path for traffic in my proposed rerouting, the major area of concern that immediately popped out was the double intersections of Murray and St. Patricks with King Edward. The problem stems from the necessity for all southbound traffic to make a left turn across King Edward. Knowing the likelihood of traffic to back up at left turn lanes, I chose this intersection to simulate, investigating whether the left turn would have an effect on my proposal, causing delays.
Figure 7: the left turn situation onto Murray.
To set up the simulation there are two key parameters to consider. The first is the demand constraints. Demand constraints are the matrix that provide the simulation the inputs for each segment of the intersection (coming from left, turning right, going straight through, etc.). The second parameter is the intersection constraints. These denote the lanes that are preferred for traffic, the stop times, the right of ways, and the behavior of the cars.
Figure 8: the demand profiles.
Figure 9: setting the intersection parameters for each approach.
Each of the vertical numbers (1-6) on the demand matrix in figure 8 represent origins for the simulated traffic, while each of the horizontal numbers represent destinations. For instance, there will be a demand of 100 between the origin of #1 and the destination of #3. In this simulation, #1 represents an origin for westbound traffic on St. Patricks St. and #3 represents an destination of a northbound exit on King Edward. This simulates traffic moving from the 417 towards Quebec.
The demand constraints in this case are proportional, rather than quantitative. If this project were seriously considered by the city, we would want to run traffic studies using real-world demand profiles, so as to connect this simulation as closely as possible with real life. However, using proportional demands will suffice for now, as long as we keep them the same for any variations on the simulation. After we are sure that the simulation is properly rendered, with correct demand profiles, we can attempt to run it. Below are some images from the simulation.
Figure 11: the rendered intersections before running the simulation
Figure 12: the simulation in Analysis view, with delay overlays on each approach
Figure 13: the rendered simulation, in which we can see the large backups, especially for those vehicles attempting to make a left turn onto Murray St.
What we see clearly in the simulation is, simply put, that if we block traffic from their current route through Rideau Street and downtown, we will create a traffic nightmare down King Edward that is even worse than what we have now. We see that southbound traffic needs to turn left onto Murray street, but is backed up through the St. Patricks St. intersection all the way up King Edward towards Quebec.
After running the simulation, we can also look at the automated analysis for the model. It shows the average delay for each approach. We see all this happening visually when we watch the simulations, but it is always good to get the data to confirm our observations; notably that this is not a good solution to rerouting the traffic.
Figure 14: the average delay times in seconds in the simulation.
The data from our traffic simulation analysis indicates a bad intersection, with large wait times and long backups. So what next? Well, one of the interesting things about Infraworks is that it allows us to customise our intersection types, and play around with simulating different intersections. Assuming we keep our demand profiles the same, we can compare delay times and flow rates to find a suitable intersection alternative to make this rerouting project work.
Roundabout
Doing some research on the different intersection types that would be good fits here, I came up with the idea of putting a urban roundabout into the second intersection area, where the left turns are taking place. Using the same demand profiles, I designed and rendered the urban roundabout.
Figure 15: the roundabout replacing the intersection
As you can see from the render, the geometry of the roundabout is a little strange. Unfortunately this is a drawback of Infraworks, where certain things are quite hard to render. When dealing with roundabouts for example, Infraworks requires that each approach and exit are perfectly opposite one another. In the King Edward to Murray St. intersection, the turn onto Murray is a little bit off from 90 degrees, which in real life wouldn't be a problem, so we will have to settle here for a rendering that does not quite line up with reality.
The same demands are used for this simulation, and generally we can reuse many of our simulation parameters for this second simulation. The simulation was run, and immediately we can see that there is minimal traffic back up in any area, thanks to the roundabout.
Figure 16: the delays. Notice a complete lack of red which denotes traffic backup.
Figure 17: the rendered traffic simulation. Notice again the lack of backup, especially in the left turning areas.
Figure 18 and 19: the difference in delay times.
We can see that when we compare the delay times from the analysis outputs, there is an obvious improvement between the two. Putting a roundabout in here is almost a perfect solution to the problems we may have created from rerouting of the traffic. So, what we have done here is indicate that the city can reroute traffic by blocking King Edward and Rideau.
Drawbacks and Conclusions
There are, of course, many drawbacks to the methodology and the project itself. Firstly, concerning the methodology, we can't guarantee that the demand profiles were inputted are accurate to the real world. For modelling the demands, I observed the intersection in-situ and referenced my observations with the inputted demand profiles, but as stated before, we would need to do a full traffic study to understand the nature of the problem. In terms of the general project, there are some concerns about large trucks and navigating urban roundabouts. In researching this issue, I came across design principles related to large traffic and roundabouts. In this case, for the state of Minnesota. “Roundabouts are designed with a truck apron, a raised section of concrete around the central island that acts as an extra lane for large vehicles and vehicles with trailers. The back wheels of the oversize vehicle can ride up on the truck apron so it may easily complete the turn, while the raised portion of concrete discourages use by smaller vehicles.” (MNDOT). In undertaking the project, it would be important to keep in mind the use by large trucks, and design accordingly, while still maintaining a slim profile.
Another issue would be the cost associated with the roundabout. In traffic rerouting projects, there needs to be clear motives for investment and action by the city. In the introduction to this piece, it is noted that the current traffic situation is dangerous to the pedestrian environment downtown, but clear data would need to be produced to show that this issue is more than an aesthetic change. Perhaps the analysis could be performed alongside the traffic studies, to find out what kind of damages and issues the current situation causes.
Finally, the whole route should be simulated, as we can not yet know exactly the kind of pressures the change will place on other portions of the new route such as the Hurdman on-ramps. Again, this can be part of a more comprehensive traffic study, showing a more detailed analysis of the issue. For the moment through, we can use the results of this investigation as an impetus to move forward with rerouting traffic in Ottawa.
Overall, This project represented a great first step in the greater project of rerouting some of the poorly designed traffic flows in the city of Ottawa. Perhaps going forward, it can help us get some ideas of better practices in the city, to eventually become a friendlier and more efficient urban area.