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

Chapter 6. General Discussion

Multiple studies were conducted to investigate the comprehension, legibility, and behavior in response to LCS and VSL ATM signs. Participants had no previous experience with the ATM signs as used in the experiments. For convenience, the current study describes the scenarios as emulating the Minnesota or Washington approaches. It is important to note that there was no intent to compare or evaluate the approaches taken in those States. Rather, the intent was to explore a range of content and contexts. The existing experimental deployments provided sources of content and contexts. Some of the sign content was novel and not used in either deployment.

A laboratory study showed that, in general, the signs were easy to interpret and would result in the actions requested (or required) by the signs. However, the following tested signs were not readily interpreted and are included in the subsequent discussion:

• Yellow arrow (flashing or steady), which was to indicate that the lane was open and to proceed with caution.
  
  
• Yellow X without text legend indicating that the lane was closed ahead. The yellow X sign with the 1 mile legend led to better performance.
  
  
• Advisory VSL signs were generally interpreted as regulatory speed limit signs.

The laboratory study also included an evaluation of sign preference by the participants. Signs that had high comprehension and were also preferred might be good candidates for deployment. These signs are the following:

• Lane Open: The participants preferred the sign as used in the Washington deployment, stating that the bolder and larger green arrow was more legible and easier to interpret from far away.
  
  
• Merge: The participants showed a preference for the streaming chevrons as deployed in Minnesota.
  
  
• Lane Closed: There was preference for the red X with the word "CLOSED" as deployed in Minnesota.

Legibility testing was done in the field using an actual CMS. This test included a subset of the ATM signs used in the laboratory experiment. Regulatory and advisory speed limit signs were tested, and the average legibility distances were approximately 1,225 ft (373 m) for all signs. These legibility distances were consistent across participants (small STDs in legibility distances were observed). The yellow X had the lowest average legibility distance at 1,040 ft (317 m). For the yellow X sign, there was a fair amount of variability in legibility distance across participants (STD of 305 ft (93 m)). The legibility distance ranged from 480 ft (146 m) to 1,250 ft (381 m), which was the maximum legibility distance because of the length of the test track. In this experiment, a response used to estimate legibility distance did not conclude until the participant finished reading the sign. The yellow X sign was shown to have poor comprehension in the laboratory experiment, and this effect was also evident in the field study. In other words, if a participant had difficulty interpreting a sign and did not respond quickly, then the estimated legibility distance was shorter. (They would be closer to the sign by the time they completed their response.)

Two driving simulator studies examined behavior in response to the ATM signs in the context of a driving situation. These studies employed signs that were shown to have the highest comprehension scores within a category of signs (e.g., lane closed signs, speed limit signs) in the laboratory study.

Side-mounted CMSs presented status information to the participants in the first simulator experiment. The eye-tracking data showed that participants fixated on these signs but not at the expense of looking at the forward roadway. Brief and direct messages appeared to be easy to understand and served as an aid in decisionmaking (e.g., lane changing behavior, reduced speed). In comparison, the second simulator experiment showed that drivers would make the correct maneuvers based on the LCS and VSL signs when no side-mounted CMS was present. The side-mounted CMS did appear to aid the drivers, but LCS and VSL signs on their own seemed to be sufficient.

The regulatory positive-contrast VSL signs were preferred in the laboratory study. In speed reduction zones, participants generally followed the posted changes in the speed limit. However, the speed reduction zones also entailed the reduction in speed for the simulated traffic. The data from the two simulator experiments showed a strong effect for the speed of the simulated (or background) traffic on participants' selected speed. In the second experiment, VSL signs were used in scenarios that entailed a lane closure (stalled vehicle or crash). In these scenarios, the VSL signs showed a reduction in the speed limit but the background traffic was not programmed to slow down. In this situation, participants initially responded to the VSL signs by slowing down but then sped up and closely matched the speed of the background traffic. These results are similar to those found by Weare, Robbins, and Blakeman where the speed of the participant driver was affected by the speed of the surrounding traffic in an ATM scenario tested in a driving simulator.⁽⁷⁾

The simulator experiments showed that the selected LCS and VLS signs resulted in effective behavior. These signs were tested in scenarios in which two left lanes, a center lane, or the tworight lanes were closed. Use of the yellow X with the 1 mile text legend resulted in the desired behavior in the simulator experiments. The laboratory test indicated that this would be a more effective sign for indicating that a lane would be closing. The streaming chevrons resulted in appropriate behavior with respect to lane change behavior. Use of the streaming chevron signs resulted in the highest comprehension and preference in the laboratory study, and they were effective in the context of the driving task in the simulator.

The manipulation of visual clutter as implemented in the driving simulator experiments had no effect on driving or eye-glance behavior. In the second simulator study, the amount of clutter on the sides of the road was manipulated by varying the number and types of buildings.

Across the two simulator experiments, the amount of information on the sign gantries varied. In the first study, the sign gantries contained only per-lane ATM signs. However, in the second study, the gantries also contained guide signs. In these experiments, the amount information on the sign gantries did not have a significant effect on driving or eye-glance behavior.

The current study 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 the high resolution and brightness of the CMS could not be modeled with a laboratory 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.