Speed Perception 2: Drivers' judgements of safety and speed on rural straight and curved roads and for different following distances

Federal Office of Road Safety - Contract Report 60 
(Performing organisation Royal Automobile Club of Victoria (RACV) Ltd)

Authors: B. Fildes, A. Leening & J. Corrigan

Full report in .pdf format [10.5MB]

Abstract

This project was a continuation of a previous study in speed perception that evaluated the role of the type of road and road width, roadside development, travel speed, driving experience and the sex of the driver on estimates of safety and travel speed (Fildes, Fletcher & Corrigan, 1987) . In the first experiment, the effect of night driving and time of testing on a driver's perception of speed on straight rural roads was assessed using the previous methodology. A validation study was performed to test whether the laboratory assessment technique was suitable for assessing speed perception at rural curves. A multifactorial experiment followed that assessed the role of the previous road, environment, and driver variables (as well as curve radius and curve direction) on drivers' estimates of safety and travel speed on flat rural curves. The last experiment was a preliminary study to see whether the speed perception methodology was also suited to testing perceptions of following distance in rural areas. The final chapter reviewed the literature on road treatments that would be suitable for use as speed perception countermeasures. A number of treatments were identified and a programme of research necessary to evaluate their effectiveness and potential road safety costs and benefits was outlined.

Executive Summary

Rural road curves tend to be over-represented in road crash statistics and perceptual cues are often claimed to be responsible.

This study set out to examine the perception of speed on horizontal and vertical curved roads in rural settings and was an extension of a previous project on the perception of speed for urban and rural straight roads (Fildes, Fletcher and Corrigan, 1987).

The Research Project

The role of type of road, roadside development, travel speed, driving experience and the sex of the driver for speed perception on rural curved roads was assessed experimentally. The effect of night vision on straight road speed perception was also evaluated.

In addition, the study examined the possibility of adapting the speed perception methodology to test unsafe driver actions on the road. A range of perceptual countermeasures for speed is also discussed, including additional research required to facilitate their use on the road.

Day and Night Vision Experiment

The first experiment involved the perception of speed on rural straight roads during the day and at night. Twelve rural straight road sites were filmed from the driving position of an automobile, simulating both normal daylight and high beam night driving conditions.

The independent variables included road type, travel speed, roadside environment, illumination level, sex of the driver, driving experience, and time of testing.

Forty eight licensed drivers were recruited and tested individually in a single laboratory sitting using both day and night stimulus scenes. Half of the subjects were tested during the day and the other half at night.

Free speed measures at each road site were taken during the day and at night. The results showed that travel speed had the strongest effect on the subjects' responses. Speeds 15 percent above the posted speed limit, were judged to be ideal, while those 15 percent below the posted speed limit were assessed to be too slow. Speed estimate errors were greater for fast than slow speeds.

These results support the previous findings of the need for moving stimulus materials when assessing the perception of speed.

Speeds on divided roads were judged to be more safe than on 2-lane roads, but travel speeds were under-estimated more on these high quality roads. Gravel road speeds were generally assessed to be quite safe, although again, travel speeds were grossly underestimated. Free speeds on the road tended to increase as the type of road improved.

Speed perceptions were generally less safe at night (and subjects made greater errors in estimating travel speed) than during the day. However, responses overall were not unsafe for either day or night scenes. There were no differences observed in free speed during either the day or night.

The level of perceived safety for daytime "spacious" sites was similarly reduced by both a "walled" roadside environment and darkness. However, trees on the side of the road had no influence at night, consistent with the lack of visual information available to drivers at this time.

Night testing had practically no influence on the subjects' responses, indicating that biological rhythms play very little part in speed perception.

The amount of driving experience had no reliable effect on either the safety of travel speed estimates, although female estimates of safe operating speed were generally less safe than male estimates.

Curvature Validation Study

A validation study was undertaken to test whether the laboratory method used for testing speed perception on straight roads was also suitable for curved roads in rural settings.

A route was selected that comprised twelve rural road curves outside the Melbourne Metropolitan area encompassing a range of spacious (open farming) and walled (heavily treed) horizontal and vertical curves.

Front seat passengers were asked to make assessments of safe operating speed and travel speed at each site. Similar responses were also made by another group of subjects in an experimental laboratory, using movie film segments of the same sites taken from the driving position of an automobile.

For horizontal curves, the pattern of results in the laboratory was similar to those collected on the road, confirming the validity of the laboratory method for testing speed perception at these sites.

However, there was less consistency for vertical curved roads, suggesting that the simple visual presentation method in the laboratory is not sufficient for testing speed perception here.

Subsequent experimentation, therefore, only assessed speed perception for horizontal rural road curves.

Horizontal Curve Experiments

A multi-factorial laboratory experiment was undertaken to assess the effect of road type, roadside environment, curve radius and direction, travel speed, driver experience and the sex of the driver on-judgements of safety and travel speed on rural road curves.

Movie film sequences were taken at 48 road curves that encompassed all the variable combinations of interest. Free speed measurements were also taken at each road curve site.

Seventy two licensed drivers with varying amounts of driving experience were recruited and tested in the laboratory.

Travel speed had the strongest effect in the safety estimate data. Speeds 15% above the speed limit were judged to be less safe than speeds 15% below the limit. Speed estimate errors were greater for slow than fast speeds.

For type of road, safety estimates varied from overly safe for divided road curves, "about right" for 2-lane road curves, to unsafe for gravel curves. Moreover, travel speeds were under-estimated much more on high quality sealed roads than on gravel curves. This finding confirms the importance of road structure and surface in the perception of speed on rural road curves.

Radius of curvature influenced both the safety and speed estimates and also interacted with type of road. Large radius curves were judged to be more safe (and travel speed was more under-estimated) than small radius curves. In addition, small radius, gravel road curves were judged to be especially unsafe.

While the range of radii do differ across these three types of roads, this finding nevertheless suggests that maintaining radius levels in the upper ranges of the standards is important for speed perception on rural curves.

There was and important interaction in the safety estimates between type of road, roadside environment and curve radius where small radius, walled, gravel curves were judged to be particularly unsafe. This result is in general accord with previous findings in curve perception and demonstrates further the need to ensure maximum sight distances in road curve design.

Curve direction had practically no effect whatsoever on both sets of data, or the speed measurements observed at the sites. This suggests that visual differences reported for different curve directions are not important in the perception of speed.

Male drivers assessed travel speed to be =re safe than females. Furthermore, novice male driver responses in particular, were less conservative than either experienced male driver or all female estimates.

Following Too Close Validation Study

A second validation study examined the possibility of adapting the speed perception laboratory method to examine the potentially unsafe driver action of "following too close to the vehicle in front".

Using a similar method to that reported earlier, a travel route was selected outside the Melbourne Metropolitan area that comprised divided, two-lane, and gravel straight rural road sites in both spacious (open farming) and walled (heavily treed) settings.

Two vehicles were used in this study. Passengers in the trailing vehicle (who were all licenced drivers) made safe distance judgements relative to the leading vehicle, and estimated their travel speed at each site for a range of different following distances and vehicle speeds.

Similar responses were made by another group of subjects in an experimental laboratory, using the same distance and speed conditions but from movie film segments taken from the driving position of the trailing vehicle.

The results in the laboratory were similar for both the safe following distance judgement and speed estimate to those observed on the road. This suggests that laboratory testing is a satisfactory means of assessing drivers' perceptions of "following too close to the vehicle in front".

An overall mean safe following distance of between 0.5sec and 2.Osec was observed from these preliminary data. Moreover, the perception of safe following distance seemed to be markedly affected by the type of road, the roadside environment, the sex of the driver, and the amount of driving experience.

There is a need for further research into following too close to the vehicle in front. In particular, a detailed study is required to systematically assess drivers' perceptions of safe following distance for a range of different road and driving conditions. The relationship between following distance and road crashes also needs to be fully assessed.

Perceptual Countermeasures

The final stage of this research program set out to review whether there were perceptual road and environment countermeasures that could be used to reduce excessive speeding in particular hazardous locations.

These measures could be expected to influence vehicle speed behaviour because they attempt to change the fundamental sensory information available to drivers. In addition, their benefits are likely to be long-term because they are unobtrusive and less likely to offend or frustrate drivers.

However, the earlier research hinted that their effectiveness would be limited if speed perception in a particularly hazardous location is overly safe to begin with (drivers did not appear to change their on-road behaviour until they perceive a particular location to be unsafe).

Eight novel road and roadside treatments were identified that could be useful as perceptual countermeasures against speeding. A planned program is described to evaluate their effectiveness and road safety benefits fully.