Impacts of Roadside Vegetation and Lane Width on Speed Management in Rural Roads

Abstract

Effective speed management in transition areas is crucial. Although numerous studies have proposed countermeasures to ensure driving safety, little research has been conducted on identifying effective and low-cost countermeasures for speed management when transitioning from rural roads to small towns. This study proposes two countermeasures: roadside vegetation and change in lane width and investigates the impact of these countermeasures on speed management performance in this context using a driving simulator experiment. Thirty participants completed eight scenarios, and countermeasures were evaluated based on stabilized speed, minimum speed, and in-town average speed. Results showed that stabilized speed and minimum speed decreased significantly in the combination of narrow lane and different vegetation designs compared to the baseline. Post-countermeasure in-town average speed didn’t decrease significantly in all scenarios. These findings suggest that roadside vegetations and narrow lane width can be effective for speed management in the transition from rural roads to small towns.

Keywords

Speed Management Driver Behavior Roadside Vegetation Lane Width Rural Road

Introduction

The national highway system, which contains interstate, arterial roads, and so on, is vital to the nation’s development (National Highway Administration, 2017). Rural roads are especially critical within this massive system as they connect people from small, often relatively isolated towns. However, while providing people with versatile travel options, these rural roads are often some of the most common places for traffic accidents. A study published by the National Highway Traffic Safety Administration (NHTSA, 2015) reveals that although traffic fatalities have been decreasing since the beginning of this century, rural traffic fatalities remain at a significant level. In 2013, 54% of all motor vehicle traffic fatalities occurred in rural areas, whereas only 19% of the U.S. population lived there. Furthermore, the fatality rate per 100 million miles traveled was 2.6 times higher in rural areas than in urban areas. Research has shown that accidents tend to occur near the entrance of small towns or cities (Ehsani, Dashtestaninejad & Khademi, 2019; Harland, Greenan & Ramirez, 2014; Casado-Sanz, Guirao & Gálvez-Pérez, 2019). These facts highlight the importance of focusing on safety issues on the transition areas in rural roads.

Speeding is one of the leading causes of traffic accidents. According to the NHTSA report (NHTSA, 2015), speeding-related issues caused 30% of fatalities on rural roads. Speed is related to the possibility and severity of traffic accidents (Elvik et al., 2019). One report divided speeding into six types and pointed out that incidental or unintentional speeding was the most frequent (Richard, Divekar & Brown, 2016). A survey also revealed that unintentional speeding was one reason for speeding among young drivers (Truelove et al., 2022). They might not be aware of speeding as their perceptions of speed were influenced by the nearby vehicles or failed to check the speedometer regularly (Truelove et al., 2022). In another survey, 60% of participants underestimated their actual driving speed while driving on the road without cameras and signs (Corbett, 2001). The study conducted by Schmidt and Tiffin found that speeding could be explained by speed adaptation (Schmidt & Tiffin, 1969). Compared to subjects who drove at high speed for a short term, those who drove at high speed for a long term were more likely to underestimate their speeds (Schmidt & Tiffin, 1969). Therefore, speed adaptation must be considered when designing roads that require drivers to transit from high-speed areas to low-speed areas to avoid speeding.

Various countermeasures have been proposed to reduce speeding-related accidents, such as speed cameras, narrowed lane width, rumble strips, and so on (Vadeby et al., 2016). Two studies explored the impact of roadside vegetation density and the distance between roadside vegetation and road on speed (Calvi, 2015; Fitzpatrick, Samuel, & Knodler, 2016). Both studies found that the impact of vegetation density was not significant on speed, and the impact of the distance between roadside vegetation and the road was significant (Calvi, 2015; Fitzpatrick, Samuel, & Knodler, 2016). One study observed that drivers drove slowly as trees were close to the road and fast as trees were far away from the road (Calvi, 2015). Similarly, another study detected that speed in small clear zone width was significantly lower than speed in large clear zone width when driving in tangents and left curves (Fitzpatrick, Samuel, & Knodler, 2016). Another study examined the effects of landscapes with different levels of greenness and complexity on driving performance (Jiang et al., 2021). They found that participants performed best in scenarios with shrubs (Jiang et al., 2021). Apart from introducing roadside vegetation, narrowing the lane width has also been confirmed as an effective countermeasure for speed management. Several studies showed that drivers drove slowly in the narrow lane scenarios and fast in the wide lane scenarios (Melman, et al., 2018; Liu, Wang & Fu, 2016; Lewis-Evans & Charlton, 2006). Godley, Triggs & Fildes found a significant reduction in the driving speed when the lane width was narrowed from 3.0m to 2.5m (Godley, Triggs & Fildes, 2004).

Previous studies have proposed countermeasures from various perspectives. However, the scenario of the small-town entrance on rural highways requires a unique solution. The existence of the town poses a limitation on the type of countermeasures that can be applied. For example, continuous vegetation is impossible due to private lands inside the town. Also, the resources for initial investment and maintenance costs are relatively limited compared to metropolitan areas and interstate highways. Towards this environment, there haven’t been sufficient studies on driving speed management during the transition from rural roads into small towns with residential areas, which is selected as the main research focus.

This study aims to identify cost-efficient and effective speed management countermeasures that can be applied near small-town entrances on rural highways. Among different options, the introduction of roadside vegetation and narrowing the lane width were taken into consideration.

Method

Apparatus

The experiment was conducted using the DriveSafety DS-600c driving simulator. The simulator is a partial Ford Focus cabin with full-width front interior, standard driver controls and active instrumentation. It renders visual imagery at 60 frames per second with horizontal field-of-view of 180 degrees. It also includes 3 configurable rear view and side mirrors that contribute to a heightened sense of realism.

Simulated Scenarios

To replicate real-life scenarios, we constructed a driving simulator environment based on the surroundings of Goodland, a town in Indiana, and the state road of U.S. 24, which runs through Goodland, as the rural highway in our simulated scenario. The scenario consists of two lanes, one for drivers traveling from west to east and the other for those traveling from east to west. Figure 1 represents the distribution of speed limit signs in one direction in this area. Drivers drive from P1 to P4. P1 is the point before entering the town with 45 mph speed limit sign. P2 is the point before entering the town with 35 mph speed limit sign. P3 is the point in the town with 35 mph speed limit sign. P4 is the point after leaving the town with 45 mph speed limit sign. Speed limits before P1 and after P4 are 55 mph.

Figure 1.

Real-world Scenario.

The simulated scenario was divided into different parts based on the speed limit: 55 mph, 45 mph, and 35 mph. Figure 2 depicts the simulated scenario, with the black line representing the 55 mph speed limit zone, the red line representing the 45 mph speed limit zone, and the blue line representing the 35 mph speed limit zone. As drivers travel from the 55 mph speed limit zone to the 35 mph speed limit zone, they will pass through a transition zone to decelerate. This transition zone covers the 45-mph speed limit zone and the 150m stretch of road beyond it, which extends from point 5 to point 3 in Figure 2. Vegetation was planted at the roadside in the transition zone.

Figure 2.

Simulated Scenario.

Design of Experiment

Three types of vegetation (short vegetation, small-spacing tall vegetation, and large-spacing tall vegetation) and two types of lane width (narrow and normal) were applied as countermeasures. Normal lane width was 3.6m and narrow lane width was 3m. In this study, high vegetation was represented by large spacing bush and low spacing bush with height equals to 2.5m. The large-spacing was 10m and the small-spacing was 5m. Low vegetation was represented by hedges with height equals to 0.7m. To determine whether countermeasures were effective or not, a baseline scenario without vegetation and with normal lane width in the transition zone was included in the experiment. This resulted in a total of eight scenarios, shown in Table 1. All scenarios were placed in daytime. A within-subject experiment was conducted.

Table 1.

Details of Scenarios.

Scenario ID	Lane Width	Vegetation
S1-Nor-NV (baseline)	Normal	None
S2-Nar-NV	Narrow	None
S3-Nor-SV	Normal	Short
S4-Nar-SV	Narrow	Short
S5-Nor-SSTV	Normal	Small Spacing Tall
S6-Nar-SSTV	Narrow	Small Spacing Tall
S7-Nor-LSTV	Normal	Large Spacing Tall
S8-Nar-LSTV	Narrow	Large Spacing Tall

Speed Management Performance Evaluation Metrics

Post-Countermeasure Town Average Speed (TAS) refers to the average speed in the area between point 3 and point 4 in Figure 2. It is expected to be small for effective countermeasure. This speed is measured after the countermeasure zone, with the intention to apply countermeasures at the entrance while reducing driving speed in the town post the transition zone.

Countermeasure Stabilized Speed (CSS) is defined as the average speed in the last 50m before point 3 (as shown in Figure 2). The effectiveness of the countermeasure is determined by the extent to which drivers decelerate and stabilize their speed during the countermeasure zone. If the stabilized speed is small, the countermeasure is effective.

Countermeasure Minimum Speed (CMS) refers to the minimum speed during the transition zone (from point 2 to point 3 in Figure 2), which is expected to be low for effective countermeasures.

Hypothesis

On the road without vegetation in the transition zone, there are two different speed limit signs in the transition zone (45mph and 35mph). The rest of the road and roadside environment in the transition zone are almost the same as the environment that is prior to the transition zone. Drivers may overlook the speed limit signs and drive into the town at a speed that is much higher than the speed limit. After planting vegetation in the transition zone, drivers will be aware of the change in the environment, pay more attention to the road and roadside environment, and reduce the possibility of missing the speed limit signs. Additionally, driver’s perceived risks would increase as trees were close to the road (Van, et al., 2018; Calvi, 2015). Under this circumstance, drivers would drive slower than in scenarios with no tree alongside the road or trees were placed far away from the road (Calvi, 2015). For cases that trees were not planted throughout the road, drivers decelerated before approaching the area with trees alongside the road and the distance between trees and the roadside was 2m (Richard & Selma, 2007). Therefore, it is supposed that the introduction of roadside vegetation can be effective in improving speed management performance.

While driving in the narrow lane, the probability of running out off the road or collision with other vehicles or roadside infrastructures will increase. In this situation, drivers are expected to drive carefully at a lower speed to make timely adjustments. Several studies have confirmed that drivers tend to drive slowly when narrowing the lane width (Melman, et al., 2018; Liu, Wang & Fu, 2016; Lewis-Evans & Charlton, 2006; Godley, Triggs & Fildes, 2004). One explanation is that drivers perceived more risks while driving in the narrow lane (Liu, Wang & Fu, 2016; Lewis-Evans & Charlton, 2006; Godley, Triggs & Fildes, 2004). Therefore, changing the normal lane to a narrow lane is supposed to improve speed management performance. Three hypotheses are as follow.

H1. Town average speed in scenarios with treatment in will be lower than in the baseline.

H2. Countermeasure stabilized speed in scenarios with treatment will be lower than in the baseline.

H3. Countermeasure minimum speed in scenarios with treatment will be lower than in the baseline.

Procedure

To recruit participants, advertisements were sent out via e-mail, IUPUI billboards, and public libraries. People who are interested in this study, have a valid driving license, and have at least one-year driving experience were selected as potential participants. All potential participants received a screening survey via e-mail and were asked to answer eight survey questions. The appropriate participants were selected based on the survey results and contacted by the group members to schedule their experiment. During the experiment, each participant first spent 10 minutes to practice driving on the driving simulator. There were 8 different driving scenarios in the experiment. Each driving scenario took 5 minutes, and the post-driving survey took 1 minute. Each participant received a $60 gift card after completing the whole experiment. This study received approval from the Indiana University Institutional Review Board (IRB protocol #: 15388). Two steps were applied to reduce learning effects:

The order of driving scenarios for each participant was randomly assigned.

To eliminate the carryover effect and consider the real-life scenario (Figure 1) that contains both westbound and eastbound lanes, the driving direction in each scenario was also randomly assigned (westbound or eastbound).

Results

Subjects

Thirty-one participants were recruited in this experiment. One participant didn’t complete the whole experiment and was dropped from further analysis. Additionally, 3 participants kept driving recklessly at abnormally fast speeds and were removed as outliers. Therefore, data based on 27 participants were used for further analysis. Among 27 participants, 13 participants were female and 14 participants were male, 14 participants were less than or equal to 25 years old, and 13 participants were more than 25 years old. Ages of participants range from 20 to 62 years (mean = 32.0, SD = 12.6).

Post-countermeasure Town Average Speed

Figure 3 displays the TAS in the baseline scenario (S1-Nor-NV) and the scenarios with treatments. According to the results, TAS was not significantly decreased after implementing treatments. Thus, H1 was not supported.

Figure 3.

TAS in Different Treatments (Error bar represents the standard error).

Countermeasure Stabilized Speed

Figure 4 shows the CSS in the baseline scenario (S1-Nor-NV) and the scenarios with treatments. CSS was significantly reduced in S2-Nar-NV (mean difference = -3.097, p-value = 0.049), S3-Nor-SV (mean difference = -4.065, p-value = 0.012), S5-Nor-SSTV (mean difference = -3.086, p-value = 0.011), S6-Nar-SSTV (mean difference = -3.619, p-value = 0.013), and S7-Nar-SSTV (mean difference = -3.041, p-value = 0.042). Therefore, H2 was supported.

Figure 4.

CSS in Different Countermeasure Treatments (Error bar represents the standard error and orange bar represents significant differences from the baseline).

Countermeasure Minimum Speed

Figure 5 represents the CMS in the baseline scenario (S1-Nor-NV) and the scenarios with treatments. CMS was significantly reduced in S2-Nar-NV (mean difference = -3.575, p-value = 0.028), S3-Nor-SV (mean difference = -3.686, p-value = 0.017), S5-Nor-SSTV (mean difference = -2.949, p-value = 0.017), and S6-Nar-LSTV (mean difference = -3.501, p-value = 0.016). Therefore, H3 was supported.

Figure 5.

CMS in Different Countermeasure Treatments (Error bar represents the standard error and orange bar represents significant differences from the baseline).

Conclusion and Discussion

The aim of this study was to investigate the impact of roadside vegetation and lane width on speed management performance. 30 subjects completed the driving simulator experiment and 27 subjects were included for analysis. Town average speed, countermeasure stabilized speed, and countermeasure minimum speed was used to evaluate the speed management performance. Paired t-test was conducted to verify whether treatments were effective. Results show that TAS was not significantly reduced after implementing all treatments. CSS was significantly reduced in S2-Nar-NV, S3-Nor-SV, S5-Nor-SSTV, S6-Nar-SSTV, and S7-Nor-LSTV. CMS was significantly reduced in S2-Nar-NV, S3-Nor-SV, S5-Nor-SSTV, and S6-Nar-SSTV. Thus, the combination of narrow lane and no vegetation, normal lane and short vegetation, normal lane and small spacing tall vegetation, narrow lane and small spacing tall vegetation, and normal lane and large spacing tall vegetation can be considered effective countermeasures.

The impact of vegetation on improving the speed management performance was confirmed by the significant reduction of CSS and CMS in different treatments. The reduction of TAS was not significant in scenarios with the introduction of vegetation, which indicated that the effect of vegetation no longer existed after the zone with vegetation. This finding was consistent with the result in (Richard & Selma, 2007). They not only detected the reduction of speed around the zone with trees but also found this impact no longer existed after passing the zone with trees (Richard & Selma, 2007).

The impact of lane width was also confirmed by the significant reduction of CSS and CMS in S2-Nar-NV and S6-Nar-SSTV, which is consistent with other researches (Liu, Wang & Fu, 2016; Lewis-Evans & Charlton, 2006). TAS was not significantly decreased in scenarios with the change from the normal lane to the narrow lane. One possible reason is CSS is significantly higher than TAS in S1_Nor_NV (mean difference = -3.087, p-value = 0.006), while CSS is not significantly higher than TAS in scenarios with treatments. The insignificant difference between CSS and TAS indicates that subjects in scenarios with treatments almost completed deceleration before entering the town and the room for deceleration in town was limited. The significant difference between CSS and TAS in the baseline scenario indicates that subjects without treatments kept decelerating after entering the town. Thus, the treatments are effective in reducing driving speed faster in the transition zone.

Another possible explanation of insignificant findings is the limited sample size. For example, the mean difference of TAS between S2-Nar-NV and S1-Nor-NV is -0.932 with a p-value equal to 0.161. Based on the following equation which is modified from (Lisa, n.d.), the required subjects should be 165 when the power is 0.80 and the significance level is 0.05. Thus, the impact of lane width on TAS might be detected after increasing the sample size.

n = \frac{{(Z_{1 - α} + Z_{1 - β})}^{2}}{μ / σ}

There are several limitations in this study. Firstly, the speed management performance was evaluated by the average speed at the end of the countermeasure and town and the minimum speed during the countermeasure. It can be also evaluated from different perspectives related to speed, such as the mean of lateral acceleration, standard deviation of speed, and so on (Liu, Wang & Fu, 2016). More speed-related metrics can be developed in the future to evaluate speed management performance. Then, the gap still remains between the simulated scenario and the real-world scenario, and the comparison between the real-road data and the data generated by the simulated environment can be made to identify the gap. Finally, the limited sample size might lead to insignificant results. Recruiting more subjects to increase the sample size for better statistical significance is necessary.

Footnotes

Acknowledgements

The work is supported by the Joint Transportation Research Program (JTRP SPR 4640) from the Indiana Department of Transportation.

ORCID iDs

Jiachen Jiang

Renran Tian

References

Federal Highway Administration. National Highway System. (2017). https://www.fhwa.dot.gov/planning/national_highway_system/ (accessed May 11, 2022).

National Highway Traffic Safety Administration. Traffic Safety Facts. (2015) https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/812181 (accessed May 11, 2022).

Ehsani Sohi

Dashtestaninejad

Khademi

. (2019). Effects of roadway and traffic characteristics on accidents frequency at city entrance zone. International Journal of Transportation Engineering, 7(2), 139-152.

Harland

K. K.

Greenan

Ramirez

(2014). Not just a rural occurrence: differences in agricultural equipment crash characteristics by rural–urban crash site and proximity to town. Accident Analysis & Prevention, 70, 8-13.

Casado-Sanz

Guirao

Gálvez-Pérez

(2019). Population ageing and rural road accidents: Analysis of accident severity in traffic crashes with older pedestrians on Spanish crosstown roads. Research in Transportation Business & Management, 30, 100377.

Elvik

Vadeby

Hels

, et al. (2019). Updated estimates of the relationship between speed and road safety at the aggregate and individual levels. Accident Analysis & Prevention, 123, 114-122.

Richard

C. M.

Divekar

Brown

J. L.

(2016). Motivations for speeding-additional data analysis (No. DOT HS 812 255). United States. National Highway Traffic Safety Administration.

Truelove

Watson-Brown

Mills

Freeman

Davey

(2022). It’s not a hard and fast rule: A qualitative investigation into factors influencing speeding among young drivers. Journal of safety research, 81, 36-44.

Corbett

(2001). Explanations for “understating” in self-reported speeding behaviour. Transportation research part F: traffic psychology and behaviour, 4(2), 133-150.

10.

Schmidt

Tiffin

(1969). Distortion of drivers’ estimates of automobile speed as a function of speed adaptation. Journal of Applied Psychology, 53(6), 536.

11.

Vadeby

Sörensen

Bolling

, et al. (2016). Towards a European guideline for speed management measures in work zones. Transportation research procedia, 14, 3426-3435.

12.

Calvi

(2015). Does roadside vegetation affect driving performance?: Driving simulator study on the effects of trees on drivers’ speed and lateral position. Transportation research record, 2518(1), 1-8.

13.

Fitzpatrick

C. D.

Samuel

Knodler

M. A

Jr . (2016). Evaluating the effect of vegetation and clear zone width on driver behavior using a driving simulator. Transportation research part F: traffic psychology and behaviour, 42, 80-89.

14.

Jiang

Chen

, et al. (2021). Moderate is optimal: A simulated driving experiment reveals freeway landscape matters for driving performance. Urban Forestry & Urban Greening, 58, 126976.

15.

Melman

Abbink

D. A.

van Paassen

M. M.

, et al. (2018). What determines drivers’ speed? A replication of three behavioural adaptation experiments in a single driving simulator study. Ergonomics, 61(7), 966-987.

16.

Liu

Wang

(2016). Effects of lane width, lane position and edge shoulder width on driving behavior in underground urban expressways: A driving simulator study. International journal of environmental research and public health, 13(10), 1010.

17.

Lewis-Evans

Charlton

S. G.

(2006). Explicit and implicit processes in behavioural adaptation to road width. Accident Analysis & Prevention, 38(3), 610-617.

18.

Van Treese

Koeser

A. K.

Fitzpatrick

G. E.

, et al. (2018). Drivers’ risk perception of roadside trees. Arboricultural Journal, 40(3), 153-161.

19.

Godley

S. T.

Triggs

T. J.

Fildes

B. N.

(2004). Perceptual lane width, wide perceptual road centre markings and driving speeds. Ergonomics, 47(3), 237-256.

20.

Van Der Horst

De Ridder

. (2007). Influence of roadside infrastructure on driving behavior: driving simulator study. Transportation Research Record, 2018(1), 36-44.

21.

Lisa Sullivan. (n.d.). Power and Sample Size Determination. https://sphweb.bumc.bu.edu/otlt/mph-modules/bs/bs704_power/bs704_power_print.html