Path Analysis - Trajectory Graphs BETA

Compass trajectory graphs visualise how long it takes for vehicles to traverse a path or intersection(s), and helps to identify where vehicles are stopped or slowed down, and queue lengths.


Efficiently managed intersections can improve journey times, and improve safety, while decreasing congestion
. Compass trajectory data consists of waypoints for connected vehicle trajectories. Each of these waypoints contains information about the location and direction the vehicle is travelling in (e.g., latitude and longitude, timestamp, speed, heading). To establish vehicle trajectory, individual waypoints are linked together to create chronological pathing information1.

Compass trajectory graphs visualise these paths. They graph how long it takes for vehicles to traverse a path or intersection(s), help to identify where vehicles are stopped or slowing down, queue lengths, and locate downstream blockages to establish which intersection is causing congestion at other sites across the network. Vehicle stoppage or queuing is indicated by horizontal sections in a trajectory line.

Users can select any section of road to understand trajectory - the tool does not have to be isolated to intersection-only use cases.

Accessing Trajectory Graphs

To access trajectory graphs, make any path selection in the Path Analysis tool. The trajectory graph will appear at the bottom of the results panel, underneath the speed and g-force graphs. Users can:

  • Filter by hours of the Day
  • Change the normalisation point (i.e., the point where all the trajectory lines converge. By default, the platform will always make the normalisation point in the middle of your selected path. You can change this by clicking on the graph to align it with another point along your path). 
  • Understand how long it took vehicles to traverse a selected path (measured in seconds)
  • Understand how long vehicles stopped for (in seconds) and where along the path they stopped.

If you have an old .ciot file that you downloaded before trajectory graphs were available, you can now update these files to include trajectory:

1. upload an existing .ciot file

2. Scroll down and press ' fetch trajectory data'

3. Trajectory graphs will load. A new .ciot file will download; replace your older .ciot file with this new version.

Interpreting Graphs

The colour of the lines on trajectory graphs indicates whether a vehicle is travelling at free-flow speeds or making several stops.

We've added a key on the side of the graph to indicate how many stops correspond to the colour of a trajectory line. The key also tells you the percentage and number of trajectories under the corresponding colours. For example, in this selection, 60% of trajectories (or 323 trajectories) had 0 stops.

Untitled design (2)

  • Green indicates free-flowing trajectories and arrival on green
  • Orange indicates one stop/split failure
  • red indicates two stops/split failures 
  • Purple indicates 3 or more stops/split failures

The flatter and more horizontal the line, the longer the time to the far side. Plateaus in a trajectory line indicate vehicle stoppage. This helps to identify where the back of the queue is. 

Below, is an example of how to interpret a trajectory line. We've removed the bulk of the trajectories to simplify this example. The normalisation point is at the middle of the intersection, indicated by the dotted horizontal line. We can see that one of these vehicles stopped for approximately 30 seconds on the approach to the intersection (highlighted in yellow).

Here's an animation of the above trajectory to demonstrate what trips along this path may look like in relation to where the vehicle is located. It also shows an example of changing the normalisation point by clicking on the graph.

Untitled design (1)

Turning on 'Advanced View' will allow you to flip the Y-axis of the trajectory graphs. To turn on advanced view, click on the profile icon at the top right-hand side of the screen and select it from the drop-down. The flip icon should appear instantly - there is no need to refresh or re-analyse your selections. 

Footnotes:
1 Saldivar-Carranza, E. D., Li, H., Mathew, J. K., Desai, J., Platte, T., Gayen, S., Sturdevant, J., Taylor, M., Fisher, C., & Bullock, D. M. (2023). Next generation traffic signal performance measures: Leveraging connected vehicle data. West Lafayette, IN: Purdue University. https://doi.org/10.5703/1288284317625