INRIX recently hosted a webinar on the use of probe data for things like queue warning. You can view a recorded version here: https://attendee.gotowebinar.com/recording/4107404171813882369 . But if you don’t have the time to watch it, I will fill you in.
The first thing to know is that one of the speakers was Todd Hartnett, Director of Business Development for ASTI. Todd described projects they now have in Maryland where they are using INRIX probe data to supplement their spot sensor data. Todd announced that ASTI has an agreement with INRIX and will now provide INRIX data as a value added service to their customers.
Todd says that probe data has several advantages. It is less expensive because you don’t have to send anyone into the field to deploy and maintain equipment. In fact, you don’t need equipment at all. ASTI has also found it useful in areas where there is not space to deploy a sensor trailer.
This may be the first time a work zone ITS provider has offered INRIX data as a standard part of their service. ASTI has apparently come up with a method to combine probe and spot data and to settle any conflicts between the two. But this example points out important differences between the two types of data that I thought we should discuss in more detail.
Simply put, probe data is usually better for monitoring overall system performance. It can show how the system, over all, is performing. It is simple, inexpensive, and easy to use. But it doesn’t provide much detail. Average speeds over a road segment are provided. How long that segment is, depends on the area and on how much you are paying INRIX for the service. INRIX does now offer a service that provides data for segments as short as 1 mile.
INRIX understands that for queue warning you must trigger those warnings quickly. And they feel segments one mile long are small enough to be effective. Certainly their system will help. But in my experience for a queue warning system to be truly effective, you need more:
1) You usually need even shorter distances between sensors (or shorter segments). Most of our work has spaced sensors an average of three-quarters of a mile apart. In some cases we have spaced them as little as a half mile apart. This is important in areas with poor sight distance such as horizontal or vertical curves and in areas with very dynamic queuing.
2) You must be able to set the locations where you measure those speeds. Probe data segments are not adjustable. You get what they give you. So a project might sit in the middle of a much larger segment. Or it might overlap 2 or more segments. Furthermore, every queue warning deployment we have provided involved some adjustment of sensor locations to catch the queue as early as possible, and report them in a consistent manner.
It also helps when you can add more sensors within work zones, and fewer in advance of them. In this way you will know not just that you have a problem, but where the source of that problem is in the work zone. Is it at the taper? Or at a lane shift in the work area? With this information you can immediately make traffic control changes and, perhaps, eliminate the cause of the queuing all together. With probe data, you won’t have that ability.
INRIX queue warning is a less expensive solution and I understand why ASTI chose to incorporate their data in their service. But when it comes to queue warning systems, you really do get what you pay for. Systems that use spot data are not normally expensive either, and they will warn drivers faster and more accurately than will a probe data system.