This article was originally published in Catalyst, November 2021.
With the improvements made to their lane-departure warning system, U of M researchers are one step closer to preventing highway crashes and deaths. In a recent project, the research team enhanced its lane-departure warning system, which uses standard GPS data rather than expensive cameras or maps—moving toward an affordable, market-ready product to warn drivers about dangerous lane drift due to drowsiness or inattention.
Using an earlier lane departure warning system (LDWS) that employs standard GPS data rather than expensive cameras or maps, Minnesota researchers have enhanced and refined the system, moving closer to an affordable product to warn drivers about dangerous lane drift and approaching curves.
ATM queue warning systems were developed and deployed on two freeways to alert motorists to queuing conditions ahead that could lead to rear-end crashes. At one test site, the prototype system substantially reduced crashes and near-crashes. At the other site, it reduced speed variances.
“The big lesson learned was that the detection method had to function quickly and display a message that was timely and accurate. This gains the trust and confidence of the motoring public,” said Brian Kary, Freeway Operations Engineer, MnDOT Metro District.
“Deploying the product of this research was not difficult. The challenge came in closing the gap to alert the drivers to slow down,” said John Hourdos, Director, Minnesota Traffic Observatory, University of Minnesota.
Kary served as the technical liaison for the study, and Hourdos was the principal investigator.
What Was the Need?
To reduce congestion and improve safety, MnDOT has deployed active traffic management (ATM) technology on two highways in the Twin Cities freeway network. The ATM system incorporates intelligent lane control signals (ILCS) placed over selected lanes at half-mile increments to warn motorists of incidents or hazards ahead. With advance warning, drivers can slow down and possibly avoid crashes.
The deployed system, however, does not specifically target the prevention of rear-end collisions, which are the most frequent type of crashes on freeways. Research has shown that rear-end collisions tend to occur during extended lines of stop-and-go traffic and at end-of-queue locations. Overhead, real-time electronic messages that warn of queuing conditions ahead can prepare motorists to reduce speed and avoid potential rear-end collisions. Such messages have the added benefit of improving mobil-ity since fewer crashes will improve traffic flow.
What Was Our Goal?
This project sought to develop and field-test two different prototypes for ATM queue warning systems. One prototype would address stop-and-go traffic and end-of-queue situations. The other would address shock waves, a crash-facilitating condition where there is a sudden change in traffic movement that causes a cascade of braking. The long-range goal of the project is to develop a unified ATM queue warning system that can be deployed at other locations within the freeway network.
What Did We Do?
Development of two prototype high-resolution ILCS warning systems began in 2014. The systems were then deployed on two high-traffic freeways in the Twin Cities: one on Interstate 35 West (I-35W) and the other on I-94. Both were still in operation in mid-2017.
The two locations have significantly different traffic conditions. On I-35W, congestion creates expanding queues that extend from the Trunk Highway 62 (TH 62) interchange to the 50th Street overpass. At the I-94 location, crashes are most likely to occur due to shock waves that can often quickly develop near the Portland Avenue overpass.
To capture traffic data, researchers used either live video from closed-circuit-camera detector stations or data from existing in-pavement loop detectors. The ILCS units dis-played the message Slow Traffic Ahead, which would direct drivers to reduce speed due to the congested lanes ahead. Other messages, such as Prepare to Stop or Traffic Ahead 10 MPH, were considered but not tested during this initial study.
A server installed at the Minnesota Traffic Observatory at the University of Minnesota archived the time and location of each queue on I-94 and measured its duration and length. This provided the data needed to develop two algorithms that can be used to develop a rear-end-collision warning system that can be installed at freeway locations where similar queuing conditions exist.
What Did We Learn?
The data collected show that warning messages delivered by the ATM system can be effective in alerting drivers to queuing conditions. The ultimate benefit is a reduction in rear-end collisions in downstream locations on the freeway.
Data recorded at the I-35W location revealed that:
Messages delivered by the ILCS system helped drivers maintain a steady speed and eliminate stop-and-go travel.
The contents of warning messages should be crafted to have an impact on all motorists. Drivers responded differently to specific messages.
Queue warning systems can be made more effective through deployment of a real-time, lane-specific ILCS system and collection of high-resolution data.
Some drivers did not always heed the first queue warning message to decrease speed, but they did slow down further along the roadway.
There was no significant difference in impact between warning messages issued during the morning peak travel period and those issued during the evening peak.
In the first three months of queue warning system operation, the crash frequency re-corded at the I-94 test site was 9.34 crashes per vehicle miles traveled (VMT) and 51.8 near-crashes. This was a 22 percent decrease from the 11.9 crashes per VMT recorded at the site in 2013 monitoring data, and a 54 percent decrease from the 111.8 near-crashes recorded there in 2013.
The research showed that to prevent potential collisions, the ATM system had to deliver messages quickly and accurately to give drivers enough time to adjust their speeds. Also, the control algorithms developed in this project can provide the queue-estimation projections needed by MnDOT and other transportation departments to enhance the effectiveness of their ATM systems.
While the deployment of the two queue warning system prototypes was a relatively cost-effective option, a longer trial period of two to three years is needed to ensure that the ATM system delivers sustainable benefits.
Imagine that you’re driving to work as usual when your smartphone announces, “Caution, you are approaching an active work zone.” You slow down and soon spot orange barrels and highway workers on the road shoulder. Thanks to a new app being developed by University of Minnesota researchers, this scenario is on its way to becoming reality.
“Drivers often rely on signs along the roadway to be cautious and slow down as they approach a work zone. However, most work-zone crashes are caused by drivers not paying attention,” says Chen-Fu Liao, senior systems engineer at the U’s Minnesota Traffic Observatory. “That’s why we are working to design and test an in-vehicle work-zone alert system that announces additional messages through the driver’s smartphone or the vehicle’s infotainment system.”
As part of the project, sponsored by MnDOT, Liao and his team investigated the use of inexpensive Bluetooth low-energy (BLE) tags to provide in-vehicle warning messages. The BLE tags were programmed to trigger spoken messages in smartphones within range of the tags, which were placed on construction barrels or lampposts ahead of a work zone.
The researchers also developed two applications for the project. First, they designed a smartphone app to trigger the audio-visual messages in vehicle-mounted smartphones entering the range of the BLE work-zone tags. A second app allows work-zone contractors to update messages associated with the BLE tags remotely, in real time, to provide information on current conditions such as workers on site, changes in traffic, or hazards in the environment.
Field tests proved the system works. “We found that while traveling at 70 miles per hour, our app is able to successfully detect a long-range BLE tag placed more than 400 feet away on a traffic barrel on the roadway shoulder,” Liao says. “We also confirmed the system works under a variety of conditions, including heavy traffic and inclement weather.”
“This was a proof of concept that showed that smartphones can receive Bluetooth signals at highway speeds and deliver messages to drivers,” says Ken Johnson, work-zone, pavement marking, and traffic devices engineer at MnDOT. “Future research will look into how we should implement and maintain a driver alert system.”
This future work includes using the results of a human factors study currently under way at the U’s HumanFIRST Laboratory to create recommendations for the in-vehicle message phrasing and structure. Then, researchers plan to conduct a pilot implementation with multiple participants to further evaluate the system’s effectiveness.
According to MnDOT, another phase of the project may investigate how to effectively maintain the BLE tag database. This phase could also investigate implementation options, such as how MnDOT can encourage drivers to download and use the app.
Two years ago, MnDOT installed a series of electronic speed limit advisory signs over Interstate 94 between Minneapolis and St. Paul. The Variable Speed Limit (VSL) system is designed to reduce congestion and help prevent crashes by recommending lower speed limits to motorists during periods of high traffic.
The new technology has worked in other places, including China and Germany. In Minnesota, a similar VSL system on I-35W reportedly had moderate benefits in reducing the total amount of congestion during the morning commute south of Minneapolis.
Although the verdict on I-94 congestion is still pending, a newly released study has found that the new system has not made a measurable impact so far on crashes in an accident-prone stretch of freeway in downtown Minneapolis. Why not?
University of Minnesota researcher John Hourdos has a few theories.
One is a simple time lag in the congestion reporting system. Another is a requirement that all lanes display the same speed limit, which he said causes confusion when only one lane is actually congested. The complexity of the I-94 commons also appears to be beyond what the VSL system was designed to do. And according to Hourdos, one of the most significant problems is the driving public simply doesn’t understand what the signs are telling them.
“People do not know what the system really does,” Hourdos said. “There hasn’t been much education on it other than a couple of news articles over the years. And when they try to decipher it on their own they get even more confused.”
The advisory speed limits are posted in response to varying traffic conditions. As vehicles approach the commons area, the system measures speeds at the bottlenecks. If the traffic slows, the system transmits a reduced advisory speed to drivers approximately 1.5 miles upstream from the location of the slow-down.
Hourdos said many motorists mistakenly believe the speed displayed on the signs is either a reflection of the speed on the current stretch of highway or an indication of the speeds on the highway ahead, rather than a suggested speed for them to follow.
The requirement to display the same speed limit on all signs also compounds the problem, Hourdos said, because when drivers see that the slowdown is only occurring in certain lanes they tend to ignore the signs altogether.
“In the lane that is congested, the real speeds drop much faster than what the VSL system can respond to, reducing the functionality of the system to the eyes of the drivers,” Hourdos said, “while on the fast-moving lanes, it seems the system has no purpose at all.”
So is the I-94 VSL system useless? Not necessarily. For one, the new study didn’t measure the system’s impact on congestion — only its ability to reduce crashes on a small portion of I-94. Moreover, the area in question, the I-94 Commons, is fairly unique, having two major bottlenecks, the highest crash rate in the state (nearly one every other day), and five hours of congestion during the afternoon rush hour alone.
“The VSL system was designed for implementation on any freeway and may not have been well-suited for the I-94 Commons area, which is a very complex corridor with high volume weaves and significant shockwave activity,” said MnDOT Freeway Operations Engineer Brian Kary.
Generally speaking, the VSL system was designed to identify slow traffic ahead of where free-flowing traffic is approaching slow or stopped traffic.
“The crash problems within the commons are caused by speed differentials between lanes and shockwave activity within the congestion,” Hourdos said. “The current VSL system was not developed to handle these types of conditions.”
MnDOT and the researchers aren’t giving up, either. A new project is starting later this year to develop and deploy a queue warning system specifically for this high-crash rate location.