Webinar 21: Iowa Department of Transportation (DOT) Snow Plow Tracking from Start to Finish

Webinar 21
Iowa Department of Transportation (DOT)
Snow Plow Tracking from Start to Finish

January 15, 2014

Summary of the Federal Highway Administration’s Quarterly Webinar: Applications of Geospatial Technologies in Transportation

These notes provide a summary of the two PowerPoint presentations discussed during the webinar and detail the question and answer session that followed both presentations.

The presentation is available upon request from the webinar speakers, Eric Abrams (Eric.Abrams@dot.iowa.gov) and John Hart (John.Hart@dot.iowa.gov).

The webinar recording is available at: https://connectdot.connectsolutions.com/p2tf0hckka2/.

Presenters

Eric Abrams
Geospatial Infrastructure and Coordination Manager
Iowa Department of Transportation (Iowa DOT)
Eric.Abrams@dot.iowa.gov

John Hart
Senior Transportation Engineer
Office of Maintenance, Iowa DOT (IADOT)
John.Hart@dot.iowa.gov

Participants

Approximately 71 participants attended the webinar.

Introduction to Presentations

Mark Sarmiento of FHWA thanked participants for joining the webinar. This webinar was the twenty-first in a quarterly series of FHWA-sponsored webinars. The series deals with the application of geographic information systems (GIS) and other geospatial technologies to transportation. This webinar focused on the Iowa Department of Transportation (Iowa DOT) snow plow tracking system, which uses technology to improve the efficiency of snow plow operations.

Presentation

Background
Eric Abrams and John Hart presented on the installation and use of an automatic vehicle location (AVL) system for Iowa DOT's snow plows. Each year Iowa DOT must deal with approximately 33 inches of snow on 9400 miles of road using 901 snow plows. In 2009, the department decided it needed to address a number of cost-control issues related to snow plowing in the State, including the high cost of salt and brine, strained budgets, and laborious record keeping. The department hoped to audit the actual amount of material used, improve consistency in response to snow events, and evaluate the impact of winter operations on pavement assets. The Iowa DOT Office of Maintenance determined that for every dollar spent on AVL systems for the snow plows, the return on investment would be $6.40, and that reducing salt use by 10 percent could save $1.4 million dollars per year.

Implementation
The main goal of the first deployment phase of the AVL system was to understand fleet movement and material (brine and salt) use. Eventually the department wanted the system to give supervisors tools to direct the fleet, decrease the amount of paperwork required from drivers, and provide the public with a better winter driving experience.

During the first phase, implemented in the winter of 2011/12, the department installed a modem with sensor inputs and a touch screen computer in a number of snow plows. Drivers logged into the system and entered in material use, time information, and road conditions, which integrated with a custom resource management system. Drivers experienced a number of issues with the systems, and Iowa DOT realized that the best AVL systems are those that require no interaction with the driver, and that instead pull data from the truck in real-time.

The following year, Iowa DOT installed modems which did not require any interaction from the drivers, on a total of 900 snow plows. Iowa DOT also changed the database architecture to make it more robust and provide the ability to add features later. The department also considered how to develop a mobile application to aid supervisors in managing their fleets.

In addition to the modems, Iowa DOT has mounted iPhones on the dashboard of a number of plows in order to view road conditions in real-time, improve situational awareness, and help assess maintenance needs. The department installed an application on the phones that allows it to function without interaction with a plow driver.

Viewing Data
Real-time data from these modems is pulled into three systems:

Internal portal, or Winter Operations Portal and Reports (WOPR): Users can search using types of map layers and can produce reports on trucks. There is also a “heads up display” for supervisors to understand what is happening to a vehicle (e.g. outside air and ground temps, types of material being used and how much, etc.) without interfacing with it.
External portal, or the WOPR Mobile site: The mobile site looks very similar to the Internal WOPR. A user can select a plow and see information about it. IADOT intends to integrate this information with traffic cameras and plow dashboard cameras in winter 2014.
Public portal, at ArcGIS online: When users click on a plow, it gives them report time, direction of travel, air and road temperature, types of materials being used, and number of plows up at any given time. The portal is a featured application on https://iowadot.maps.arcgis.com/home/index.html.

Demonstration
IADOT gave a demonstration of the AVL system. IADOT's maintenance page links to the snow plow locations interface, where the agency is hoping to soon display some road conditions information. On the site, the public is given two options: advanced and basic plow sites. The advanced site shows active plows and cameras, with a direct link to resources and the ability to turn on and off map layers. The basic site shows the same information, but does not allow the user to customize what is displayed. In the future, IADOT would like to provide more robust winter information so that the public can make better traveling decisions based on road conditions.

In the WOPR, a user can hover over a plow truck display truck and obtain information such as inventory, materials, speed, rates, plow status, and potentially odometer and engine hours. This information can help IADOT determine when a truck needs to return to the shop for maintenance. WOPR also displays crumb trailing, which shows where trucks have been over last six hours. This is helpful during storms to show plow density and coverage.

Question and Answer Session

Are there two data plans per truck with the camera—one for the iPhone and one for the modem?
The phone is not on the same data plan as the modem. Eventually the system may move to that configuration, so that data can be integrated and funneled through one data plan.

Are data saved when the truck is out of cellular coverage and transmitted when a connection is reestablished?
Yes. A few days' worth of AVL data can be saved in the unit at one time. Then the data will transmit back to the database once it is in cell range.

Can you go back in time and pull up data on a particular event?
Yes, we keep all the data from each year. Last winter we harvested and stored 60 million pings. We have the ability to go back in half-hour increments to provide time-lapsed assessments of storms and overlay traffic speeds, material use, performance measures, etc. to conduct analyses.

Have you verified the accuracy of the data from the spreader controllers?
We believe that the controllers are very accurate. This has been demonstrated through field verification. We are currently concerned with the accuracy of the algorithms. The truck pings every minute. How do the minute-by-minute data compare with the data that comes out of the spreader controller? This question will be explored further in winter 2014.

What is the data management plan, considering the large amount of data that are downloaded in real time? Are some of the data archived, etc.?
When it comes to the AVL system, we have a 5 MB data plan per truck. The data are not very complex. The phones have an unlimited data plan, but each picture (taken approximately every five minutes) is 500 KB, which is still minimal. Additionally, the trucks do not run constantly. We do not anticipate any issues related to data storage.

Are you tracking when the plows are up and when they are down?
Yes. The mercury switch tells us whether the system is up or down. However, it is unreliable because it bounces around as the truck drives and turns it off sometimes. This is not the best way to track it.

Can you get a report for a specific truck over a specific time period or just the real-time data?
We can do both. We can go back in time and pull data for an individual truck at specific times to see exactly what it was doing by developing a SQL query or selecting it in the WOPR.

Have cost savings been verified? If so, how was it done?
Not yet. At this point, we have only looked at general trends of salt usage and tracking material usage. No in-depth study or comparison has been conducted. The first two years were just a trial, but we are now beginning to do some deeper analysis that can be used to drive decisions.

Do the crews put the data into a Maintenance Management System where you could record labor, equipment, etc.?
That is one of the goals. The IT department is working on pulling material usage and the time that the truck was out, but it is only done manually at this point. We do not have an automated process to sync the materials used with other data. Eventually we want to bring everything into the system.

Can we export the data in .csv or Excel spreadsheets?
Yes, there are lots of options for exporting. You can export to KML (Keyhole Markup Language) or access the data via ESRI rest endpoint. You can even hook up to Oracle with GIS software.

How do you ensure the accuracy of road conditions assessments compared to the 511 website? If the public sees a blizzard in the dashboard camera feed, how can you be sure that this is in agreement with the 511 report?
We have switched gears in how we are doing condition reporting. It used to be handled by state patrol, but it is now the responsibility of maintenance supervisors. They individually enter road conditions, and that is what is displayed on the 511 website. They are using visual assessments and driving the roads themselves. Eventually, the approach may be to combine visual assessments by the maintenance supervisors with output from the dashboard cameras to assess road condition for the 511 website.

On one of the screenshots, it looked like there was a real time weather layer on the interface. Is that correct?
Yes, there is a real-time radar from Iowa State University, which has a five to ten minute refresh rate.

You mentioned earlier that you set up a cloud server to handle internal information. Did you have typical client server set up before? How does using the cloud compare to the typical client server set up?
Yes, we did have a typical client server before. We will soon find out how it compares. We decided to use a cloud-based server because we no idea what the usage was going to be. We did not want to roll out the interface to the public and then find that the site isn't working. We also chose cloud-based for the storage of pictures and KML files. However, we can't get an updated KML up on ArcGIS Online because there is no application programming interface (API) for that. We went with Microsoft cloud storage which is referenced on ArcGIS Online.