Active Traffic Management: Comprehension, Legibility, Distance, and Motorist Behavior in Response to Selected Variable Speed Limit and Lane Control Signing

Executive Summary

Active traffic management (ATM) incorporates a collection of strategies allowing the dynamic management of recurrent and nonrecurrent congestion based on prevailing traffic conditions. These strategies help to increase peak capacity, smooth traffic flows, and increase safety on busy, major highways. Some popular approaches include variable speed limits (VSLs), lane control signs (LCSs), and hard-shoulder running, which are all controlled by overhead changeable message signs (CMSs).

This report describes four studies (one in a laboratory setting, one in a field setting, and two in a driving simulator) in which two particular ATM approaches were researched: VSLs and LCSs. The field study was conducted at a drag strip in Manassas, VA. All laboratory and simulator studies were conducted at the Federal Highway Administration's (FHWA) Turner-Fairbank Highway Research Center. Across these latter three studies, the following five scenarios were used to investigate participant response under various driving situations. Scenarios were presented on roadways with four lanes in the direction of travel, where lane 1 was the leftmost lane and lane 4 was the rightmost lane.

• Resting condition for the signs (normal operations for all lanes).
• Stalled vehicle closing one lane (lane 3).
• Vehicle crash closing two left lanes (lanes 1 and 2).
• Vehicle crash closing two right lanes (lanes 3 and 4; with exit ramp open).
• Message in the event of slow-moving traffic in all lanes (speed reduction).

The laboratory study examined how well drivers interpreted different VSL and LCS configurations, symbols, and messages as a function of scenarios. In addition, the comprehension of the different alternatives for the signs was tested (e.g., regulatory versus advisory speed limit signs, different signs for merge). Participants also rated their preference for alternative ATM signs. Sign stimuli were created using actual sign specifications and other information provided by the Minnesota and Washington transportation departments and the sign manufacturer for the Washington deployment. Individual sign types included lane open, lane open with caution, lane closed, lane closed ahead, merge, and VSLs.

The results of the laboratory study showed that participants frequently interpreted the ATM signs correctly as the signs were presented in sequence for a given scenario. Participants in the study had no previous experience with these types of signs, suggesting the ATM signs were, for the most part, intuitive. Errors included interpreting advisory VSL signs as regulatory speed limit signs, incorrectly interpreting green overhead guide signs, misinterpreting the lane closed ahead sign with a legend, and confusing the meaning of the lane open with caution options (both static and flashing). Participants were challenged in the scenario in which the two right lanes were closed except for the exit. One participant stated, "How can the lanes be open and closed at the same time?" These results were based on static testing in which participants only needed to pay attention to the signs. There was no driving task or other additional workload present in the test situation.

A subset of the ATM signs used in the laboratory study were selected for further field testing. This test was for a large number of CMS messages in which the ATM signs were a small subset. The principal focus of the field test was to obtain legibility distances for the signs. Data were collected on a closed course while participants drove an instrumented vehicle equipped with a dashboard-mounted eye-tracking system. Audio recordings were made to document participant response and comprehension.

The mean legibility distances for the speed limit signs were close to the maximum legibility distance in this test (1,250 ft (381 m)). The lane closed ahead sign had the lowest mean legibility distance (1,040 ft (317 m)) and the highest standard deviation (STD). The greater legibility distance and STD in responding to this sign was in part a result of the method used to obtain a response. A response did not conclude until the participant completed reading the message. Thus, if a participant struggled in interpreting the true meaning of the sign (i.e., the lane was closing in the near future), then the response duration was longer. In turn, the legibility distance decreased. A factor in the design of signs and the letter height used is the available time to read a sign at highway speeds. The lane closed ahead sign had an average legibility distance of 1,040 ft (317m) at a speed of 65 mi/h (105 km/h); the reading time was approximately 11 s, which was greater than the design guideline minimum of 8 s. The ATM signs using the new CMS displays with high resolution resulted in legibility distances exceeding design guidelines for non-light-emitting diode (LED) signs (e.g., highway guide signs).

The next study used a driving simulator to examine driver decisionmaking in a dynamic environment. The study employed the same scenarios that were used in the laboratory study (the speed reduction scenario was shown twice) and examined drivers' responses (behavioral and eye-gaze movements) to and compliance with LCS and VSL signs.

Participants generally followed the directions on the ATM signs. In the speed reduction scenario, VSL signs showed lower speed limits every 0.5 mi (0.8 km). Participants decreased their speeds about 328 ft (100 m) after passing the signs. Although drivers slowed down in response to the VSL signs, they were traveling above the speed limit by an average of 10 mi/h (16 km/h). Eye‑gaze behavior showed participants looked at the VSL signs about two times during these scenarios. Mean fixation durations ranged from 0.3 to 0.5 s and decreased significantly from the first to the second viewing of the VSL signs.

For the stalled vehicle scenario, participants exited their travel lane upon encountering the stalled car; no participant collided with the vehicle. Most of the participants (about 70 percent) exited their lane well in advance of when they were required to exit the lane. From a safety perspective, this may prove to be an effective behavior. However, early exiting of the lane by a large volume of traffic may not be optimal behavior from a traffic flow perspective. On average, participants slowed to about 5 mi/h (8 km/h) below the speed limit while approaching the stalled vehicle. The mean fixation duration to the LCS for this scenario was about 0.51 s, which was significantly longer compared with the other scenarios. Participants were able to obtain information from the signs without long (greater than 2 s) fixations.

When the two left lanes were closed because of a crash, most of the drivers remained in their lane or moved to the right lane, both of which would be correct responses. Speed behavior was similar to that observed for the stalled vehicle scenario. Average fixation duration to the LCS was about 0.4 s, which was shorter than for the stalled vehicle scenario but similar to the other scenarios. Participants were able to obtain information from the signs without long (greater than 2 s) fixations.

The scenario with the two right lanes closed owing to a crash was a bit challenging. Drivers were instructed at the start of the simulation to exit at Holt Ave (the exit in this scenario). Results showed the majority (74 percent) of the participants successfully exited at Holt Ave according to instructions. Drivers slowed (less than 5 mi/h (8 km/h) until they approached the exit and then significantly decreased their speeds as they neared the exit.

The second simulation study used the same scenarios as the previous laboratory and simulation studies. In addition, two different levels of roadway clutter (low and moderate) were included on the sides of the road. The focus of this study was driver behavior and decisionmaking under different scenarios.

Driver exit-taking behavior at Holt Ave was reviewed to determine whether the amount of visual information on the sides of the road had an effect on participant response. Results showed there was no difference in the exit-taking behavior under low and moderate levels of clutter. Overall, about 63 percent of the participants correctly exited at Holt Ave.

In the speed reduction scenario, the VSL signs displayed reduced speed limits every 0.5 mi (0.8km). Participants decreased their speeds after they encountered the signs. Behavioral results of this second simulation experiment were similar to the results of the first. Eye-gaze data showed participants looked at the VSL signs an average of six times during these scenarios, and fixation durations averaged about 0.33 s. Compared with the first simulation study, there were more fixations on the VSL signs, but they were of a shorter duration.

For the stalled vehicle scenario, participants exited their travel lane upon encountering the stalled car; no participant collided with the vehicle. The majority of the participants (about 61 percent) exited their lane well in advance of when they were requested to do so. In this experiment, participants were shown VSL signs reducing the speed limit as they approached the stalled vehicle. Drivers responded to the VSL signs but were still traveling about 5 mi/h (8 km/h) above the new posted speed limits. These results were similar to those found in the previous simulation experiment. The mean fixation duration on the LCS for this scenario was about 0.33s. However, there were multiple fixations on these signs (on average about eight fixations). This was different from the first simulation experiment where there were fewer, but longer, fixations on the LCS. Participants obtained information from the signs without long (greater than 2 s) fixations.

When the two left lanes were closed because of a crash, most of the drivers remained in their lane or moved to the right lane, both of which would be correct responses. Speed behavior was similar to that observed for the stalled vehicle scenario. Participants slowed in response to the speed limit decrease shown on the VSL sign but were traveling about 5 mi/h (8 km/h) above the new posted speed limit. The average fixation duration on the LCS was about 0.33 s, and there was an average of 7.5 fixations on the ATM signs. Fixations were of short duration but frequent, similar to what was observed in the stalled vehicle scenario. Participants obtained information from the signs without long (greater than 2 s) fixations.

The scenario with two right lanes closed due to a crash was a bit challenging. Drivers were instructed at the start of the simulation to exit at Holt Ave. Results showed the majority (63percent) of the participants successfully exited at Holt Ave according to instructions. In this experiment, VSL signs were used to reduce the speed limit during the scenario. Data suggested participants did respond to the reduced speed limit and drove approximately 10 mi/h (16 km/h) above the new speed limits. The mean fixation duration on the ATM signs in this scenario was 0.33 s. However, the number of fixations on the ATM signs in this scenario was significantly greater (about 10 fixations on average) than for all of the other scenarios. Participants had many fixations on the ATM signs but they were of short duration.

The current report used a range of approaches to test different sign options for ATM. Numerous signs were screened using laboratory methods suitable for estimating sign comprehension and preference. In addition, field testing used an actual CMS for estimation of legibility distance because high resolution and brightness of the CMS could not be modeled with a laboratory liquid crystal display (LCD). Use of a subset of ATM signs in the field allowed accurate estimation of the legibility distance of these signs. A driving simulator was used to evaluate different sign options under various driving conditions, which allowed the evaluation of responses to the signs in a dynamic environment. The combination of approaches used resulted in assessment of the signs' comprehension, legibility distance, and effect on driver decisionmaking.