In the spirit of being open, we’re gearing up for an upcoming series of blog posts that focus on clarifying some of the more vague – and ultimately confusing – aspects of our industry. We’ve already argued that the building of smart cities should start at the intersection, and now we want to dive into some other topics we see people using (and abusing) day to day.
One metric that traffic engineers use to design safer streets and better cities for citizens is travel time. While industry terms and methods for gathering travel time data vary, we’re here to clear up some of the confusion around the different terms people use to muddy the water.
Travel time is a widely used metric that is crucial to the success of many types of traffic projects, including signal retiming and congestion management projects. It’s one of the metrics used by traffic engineers to evaluate the performance of different transportation facilities such as urban arterials, highways, and freeways. Travel time is typically used to conduct before-and-after studies, speed studies, and congestion studies. This type of data is also used to quantify the effects of changes – such as construction – to roadways on traffic network performance, and to calibrate simulation models.
There are two main factors that affect travel time accuracy: the outlier filtering algorithm and sample size. This post will focus on what we think you should consider before choosing the method by which to conduct an accurate travel time study, as it relates to sample size.
In order to ensure an accurate representation of real-world traffic conditions, you need a sufficient number of data points. That’s because any one given trip may not be representative. Travel time, especially in urban arterials, is subject to high variability because of the interrupted nature of traffic flow. An accurate average travel time measurement cannot be guaranteed if a sufficient number of data points are not collected.
The best way to get a lot of data points is by having both high capture and high match rates. While some vendors may attempt to equate the two, it’s important to understand the difference between them. This will help ensure you have the right data to make informed decisions on which technology would best produce data you can trust to meet FHWA guidelines for your next travel time study.
Interested in learning more about sample size and travel time? Read our Technology Showcase.
Match rate vs capture rate: what’s the difference?
A capture rate is the ratio of the number of unique MAC addresses detected at one location to the total number of vehicles that traveled past that location in a given time period. For instance, if a car travels through intersection A and the MAC address is picked up, it counts as a single capture.
A match rate is defined as the ratio of the number of trips detected between intersections A and B, to the total number of trips that actually occurred between the points. This is a more difficult number to measure, because of the expense and difficulty of obtaining ground truth data regarding the total number of vehicles that originated from A and ended up at B.
When capturing travel time, devices will inevitably pick up signals coming from unwanted sources. From parked vehicles or nearby stationary devices, to someone walking by with a connected device, there are many ways these unneeded data points can creep into the data. Typically this data is left to either contaminate your sample, or for you to manually find and remove. While most solutions either give you raw data or simple reports, it’s important to work with data that has been filtered to remove outlier events. Additionally, these filters should have the ability to be fine tuned, allowing you to more precisely achieve the result you’re looking for.
All of this, plus the fact that there are often opportunities for traffic to enter and exit the travel time corridor, means that getting a ground source of truth for your traffic study is almost impossible. But with Miovision, you can access the collected video, which provides you with an accurate approximation of the ground truth data, so you can analyze the capture rates you’re achieving with our solution.
How does Miovision do it?
Match rates and capture rates can be calculated for any traffic detection device, including those using WiFi and Bluetooth communication standards. In our case, the detection device is a Miovision Scout that sits at the roadside collecting MAC addresses from passing WiFi-connected devices. WiFi probes tend to gather MAC addresses at a high rate, because people are more likely to keep WiFi enabled on their smartphone at all times. By keeping WiFi enabled continually, smartphones are able to search for a connection to a free WiFi hotspot. On the other hand, many smartphone users will switch off ‘discoverable mode’ to save battery life, or to ensure their device isn’t constantly searching to establish a connection with another Bluetooth-enabled device.
In the end, the best way to achieve accurate travel time studies, is to start with trustworthy data. Miovision DataLink provides traffic engineers with match rates that exceed FHWA guidelines, which are essential to accurate travel time reporting. With the right technology in place, you can make informed decisions for your city and make a difference in the lives of citizens, one intersection at a time.
Read our Technology Showcase to learn more about match and capture rates.
If you enjoyed this blog post, you may also like: