Scholarly article on topic 'Platooning on Two-lane Two-way Highways: An Empirical Investigation'

Platooning on Two-lane Two-way Highways: An Empirical Investigation Academic research paper on "Civil engineering"

CC BY-NC-ND
0
0
Share paper
OECD Field of science
Keywords
{"two-lane highways" / platooning / car-following / passing}

Abstract of research paper on Civil engineering, author of scientific article — Ahmed Al-Kaisy, Casey Durbin

Abstract This paper presents an empirical investigation into platooning on two-lane two-way highways. A total of six data sets from three study sites in the state of Montana were used in this investigation. Field data included individual vehicle speed, time gap, and vehicle classification. The study confirmed that interaction between successive vehicles on the same lane in the traffic stream generally diminishes beyond a time headway threshold value that fell in the range of 5-7seconds. Also, the study revealed that very short headways are more associated with aggressive driving and higher speeds and that the amount of impedance to traffic is proportional to the size of platoon.

Academic research paper on topic "Platooning on Two-lane Two-way Highways: An Empirical Investigation"

Available online at www.sciencedirect.com

ScienceDirect

Procedia Social and Behavioral Sciences 16 (2011) 329-339

6th International Symposium on Highway Capacity and Quality of Service Stockholm, Sweden June 28 - July 1, 2011

Platooning on Two-lane Two-way Highways: an Empirical

Investigation

Ahmed Al-Kaisya, Casey Durbinb

aMontana State University, Civil Engineering Department, Bozeman, Montana, USA 59715 b Wickens Construction, Inc, Lewistown, Montana, USA 59457

Abstract

This paper presents an empirical investigation into platooning on two-lane two-way highways. A total of six data sets from three study sites in the state of Montana were used in this investigation. Field data included individual vehicle speed, time gap, and vehicle classification. The study confirmed that interaction between successive vehicles on the same lane in the traffic stream generally diminishes beyond a time headway threshold value that fell in the range of 5-7 seconds. Also, the study revealed that very short headways are more associated with aggressive driving and higher speeds and that the amount of impedance to traffic is proportional to the size of platoon.

© 2011 Published by Elsevier Ltd.

Keywords: two-lane highways, platooning, car-following, passing

1. Introduction

Platooning is an important phenomenon on two-lane highways that has serious implications on traffic operations. Specifically, the quality of service on two-lane highways is directly related to the platooning phenomenon and is estimated using two surrogate measures; the percent-time-spent-following (PTSF) and average travel speed (TRB, 2000). Platooning, in turn, is a function of the amount of passing opportunities available in each direction of travel which is largely dependent on headway and speed distributions for a given volume of traffic. Platooning is also important on two-lane highways from safety perspective. Drivers who are constrained by slow moving platoons and lack of passing opportunities may become frustrated, and therefore tend to accept smaller gaps in the opposing traffic to perform risky passing maneuvers, many of which are aborted which result in unsafe traffic conditions particularly at higher speeds.

1877-0428 © 2011 Published by Elsevier Ltd. doi:10.1016/j.sbspro.2011.04.454

In general, the amount of platooning on two lane highways is known to be a function of traffic volume in the same direction of travel, opposing traffic volume, percent no-passing zones, and speed variation within the traffic stream. Expectedly, those are mostly the same factors that would primarily determine the need for overtaking and the amount of passing opportunities for a specific traffic stream.

This paper presents an investigation into the platooning phenomenon on two-lane two-way highways and provides analyses that will help in developing a better understanding of traffic operation on two-lane highways.

2. Background

While there has been extensive research on the platooning phenomenon at signalized intersections, platooning on two-lane highways has received only limited attention. This is despite the fact that platooning is a major determinant of traffic operations and safety on two-lane highways. The Highway Capacity Manual (HCM) refers to the "formation of platoons" as an important phenomenon in determining traffic performance on two-lane highways (TRB, 2000). An important performance measure on two-lane highways, percent-time-spent-following, used by the current HCM is largely related to the amount of vehicular platoons. A few studies on the capacity and quality of service on two-lane highways, passing lanes, and passing maneuvers qualitatively described this important phenomenon in the context of describing patterns of vehicular interactions without examining the physical attributes of platooning (Gattis et al., 1997; Guell and Virkler, 1988; Luttinen 2001; Roess, 1984). A few older studies attempted empirical measurement of platooning on two lane highways. More notably, some of those studies investigated the relationship between traffic flow and platoon size using "critical" headway in defining platooned vehicles such as the study by Morrall and Werner (1982) and an earlier study by Underwood (McLean, 1989). A very interesting overview of those studies is presented by Mclean (1989) who concluded that "empirical measurements of platooning have been limited to specific sites and there have been no systematic measurements of two-lane flow platooning or headways." No recent empirical research on two-lane highway platooning has been identified in the literature with the exception of a study by Sadeghhosseini and Benekohal (1997) that investigated the relationship between traffic volume and platooning using headway definition of platoons.

3. Definition of platooning on two-lane highways

It is fair to state that the term "platooning" has been used in the context of traffic operations on two-lane highways only subjectively. Specifically, there is no clear definition as to what constitutes a platoon on a two-lane highway or how to measure this vehicular behavior in a more quantitative manner. The Highway Capacity Manual (TRB, 2000) defines a platoon as "a group of vehicles or pedestrians traveling together as a group, either voluntarily or involuntarily because of signal control, geometrics, or other factors." The HCM, however, does not provide further details or interpretation of "traveling together as a group" in order to identify platooned vehicles. Specifically, the HCM does not specify how close those vehicles need to be, nor does it suggest a particular number of vehicles that constitute a "group." Researchers have used the general definition of the HCM without a consensus on the quantitative attributes of a vehicular platoon on a specific type of highway (Arasan and Kashani, 2003; Gattis et al., 1997). A study by Gaur and Mirchandani (2001) states that "an observer may consider two or more vehicles to be in a platoon either because of their closeness or because of their relative distance from other vehicles on the link." While two or more vehicles sounds a logical consideration for platoon size, the study did not specify how close successive vehicles in the traffic stream must be in order to be identified as platooned vehicles.

4. Research motive

The importance of the platooning phenomenon in determining traffic performance and safety on two-lane highways, and the limited research that has been done in this regard were the main motives behind the current research. In order to better plan, manage, and operate two-lane highways and to advance the current analytical procedures, a better understanding of the platooning phenomenon is needed.

5. Data collection and processing

5.1. Study Sites

Three study sites in the state of Montana were selected for this study. In the selection of study sites an attempt was made, as much as possible, to include sites with a wide range of traffic conditions. Also, study sites were selected to represent uninterrupted flow conditions away from the influence of driveways, intersections, or other access points that may have implications on vehicle speeds and behavior. A description of these study sites is provided in this section.

Study Site 1 - Jack Rabbit Lane

This study site is located south of Belgrade, Montana on Jackrabbit lane which is a rural two-lane highway that runs north-south through the city of Belgrade up to Four Corners, a small town just a few miles west of Bozeman, Montana. This segment of jackrabbit lane runs between the interstate I-90 to the North and state route 191 to the south in a relatively flat terrain. There are frequent access points to roadside developments towards the north and south ends of this segment. The study site location was chosen near the middle of this 10-mile segment in an attempt to avoid the effect of traffic control devices or other interruptions on field data. Traffic level at the study site is relatively high with traffic mainly consisting of local drivers commuting to work in the morning and the afternoon peak hours, and more non-local traffic that is destined to Big Sky and West Yellowstone areas during off-peak hours. While all study sites have relatively high traffic volumes, this site witnesses higher traffic levels compared with the other two study sites.

Study Site 2 - Highway 287 South

Highway 287 is a two-lane state route that connects interstate I-90 near Three Forks, Montana to the south with Interstate I-15 at the capital city of Montana, Helena to the north. This site is located on the segment of Highway 287 which connects the town of Three Forks, Montana to the south and the town of Townsend, Montana to the north. This particular segment extends for roughly 33 miles in a rural setting with much of the roadway traversing rolling terrain. This site has relatively high traffic volume as it serves traffic traveling between the capital city Helena and Interstate 90. The location of study site is around 7 miles to the north of I-90 on a highway section that is mostly in flat terrain.

Study Site 3 - Highway 287 North

This site is located on the segment of Highway 287 that connects the town of Townsend, Montana to the south with the capital city Helena to the north. This is a 36-mile segment that runs in a rural setting with much of the roadway traversing flat or hilly terrain. The location of study site is around five miles north of Townsend on a highway section that is considered level. Traffic increases at this segment of Highway 287 during the summer due to its location near three recreational lakes, which made it a promising location to analyze unique traffic mixes and traffic peaking characteristics. There are passing lanes around two miles to the north of this location in the two directions of travel.

5.2. Field Data

Two data sets from each study site were collected using pneumatic tube traffic counters (one set in each direction of travel). Speed, time gap, length and vehicle classification for each individual arrival were recorded for the duration of data collection. A total of 236 hours of traffic data was collected from all study sites. Description of field data is provided in Table 1. Vehicle length and time gap information were used to estimate time headway for individual vehicles in the traffic stream.

Table 1 Description of Field Data

Northbound

Southbound

Study Site

Duration (hr)

Total Vehicle Count (veh)

Duration (hr)

Total Vehicle Count (veh)

Highway 287 South July 1, 2005 - July 81.90 8393 81.90 7960

4, 2005

Highway 287 North July 1, 2005 - July 2, 2005 20.75 2672 20.75 2874

Jackrabbit Lane July 31, 2005 -August 1, 2005 15.50 2128 15.50 3491

6. Study results

In this research, a vehicle is considered in a platoon if it is, voluntarily or involuntarily, in car-following mode, i.e. maintaining an appropriate headway from the lead vehicle. This definition ensures that there is interaction between successive vehicles that are part of the same vehicular platoon. Based on this definition, it sounds logical to associate platooning phenomenon on two-lane highways with either or both of the following:

1. A headway that is equal to or less than the comfortable safe headway as perceived by the driver of the following vehicle

2. An actual vehicle speed that is lower than driver's desired speed

While the above definition of platooning may sound clear and straightforward, it is not easy to measure in the field. This is mainly due to two important issues:

1. Appropriate headway and desired speed are both traffic parameters that are stochastic in nature, i.e. vary based on driver and vehicle characteristics

2. Some vehicles in platoons may be in car following mode, yet their speeds are not much different from their desired speed. Those are mostly vehicles that are referred in the literature as being "voluntarily" in platoons or "happy to follow" lead vehicles.

To better understand the platooning phenomenon in the context of the above discussion, it is essential to examine the relationship between time headway and average travel speed on two-lane, two-way, highways. To address the stochastic nature of headway and speed measurements, an aggregate approach was utilized to perform the analyses in this research.

In this research, any two or more consecutive vehicles traveling in a group (identified by short headways) would be considered as a platoon. The first vehicle in the group is referred to as platoon leader, while all other vehicles traveling behind platoon leader are simply referred to as platooned vehicles.

6.1. Relationship between Travel Speed and Time Headway

This relationship is very important in understanding the dynamic of platoon formation and the fact that vehicles in platoons may be traveling at less than their desired speeds due to lack of passing opportunities. It is logical to expect an association between the proportion of those vehicles impeded by platoon formation and small headways in the traffic stream. This relationship is also essential and important in identifying the headway size beyond which interaction between successive vehicles in the traffic stream becomes negligible. In other words, the relationship can be used in identifying vehicles that are outside the influence of vehicular platoons.

Data from study sites was analyzed so that mean travel speed was calculated for all vehicles in the traffic stream with headways that are equal or greater than a specific threshold value. Results from this analysis are exhibited in Figure 1. The general trend shown in this figure is the increase in mean travel speed as time headway threshold increases. However, the rate of increase in mean travel speed decreases steadily until curves flatten out when headway threshold exceeds certain value in almost all cases with the exception of HWY 287 North - SB data that showed a relatively different pattern.. This leveling out of curves clearly corresponded to headway threshold values of 5-7 seconds. The limiting value of this almost asymptotic curve should be close or equal to the mean desired speed of vehicles that are outside the influence of platoons. At small time headway threshold values, mean speed represents a traffic stream with larger amount of platooned vehicles and thus higher level of interaction between successive vehicles, i.e. more vehicles in car-following mode. On the other hand, at greater time-headway threshold values mean speed represents vehicles with low level of interaction and more vehicles traveling outside platoons at

their desired speeds, i.e. not in car-following mode. Data from the site HWY 287 North Southbound were different in that speed is less sensitive to time headway compared with other data sets (notice the range of change in speeds). A 1.5-mile long passing lane exists at around two miles upstream of this study site. The presence of a passing lane in a close proximity from the study site results in breaking up platoons and more vehicles speeding up to their desired speeds (upon leaving platoons) with relatively short time headways. This is the main reason why the fundamental relationship between short headways and vehicles being impeded by slower-moving vehicles is not very clear at this particular site.

HWY 287 N - NB

Headway Threshold (sec)

"g 64.5 -a

— 64 -

H 63.5 -¡a

S 63 -

HWY 287 N - SB

4 6 8 10

Headway Threshold (sec)

HWY 287 S - NB

Headway Threshold (sec)

Jackrabbit - NB

Headway Threshold (sec)

60.5 -| 60 -59.5 -59 -58.5 -58 -

HWY 287 S - SB

s 2 4 6 8 Headway Threshold (sec) 10

Jackrabbit - SB

Headway Threshold (sec)

Figure 1: Relationship between Mean Travel Speed and Headways Being Equal or Greater than a Specific Threshold Value

To better understand the relationship between mean speed and headway threshold, the correlation coefficient was found and included in Table 1 along with the equations for the fitted curves shown in Figure 1. In general, there is a very strong relationship between mean travel speed and time headway threshold in most study sites investigated. In five of the study sites, the correlation coefficient exceeded 0.8. The only exception was the data set from Highway 287 North Southbound where the correlation coefficient is only around 0.24. The presence of a passing lane upstream of this study site is believed to be the source of discrepancy as discussed earlier in this section. In regards to curve fitting and regression equations, it was found that a third-degree polynomial curve provided an excellent fit to field observations as confirmed by the very high coefficient of determination (R2) for all curves exhibited in Figure 1.

Average travel speed is plotted against individual time headways at all study sites to gain more insights into this fundamental relationship, as shown in Figure 2. Time headways were grouped into ranges of one second in duration. This figure confirms the main finding from Figure 1 that the increase in speeds is more notable at short headways and relatively diminishes when time headway exceeds a certain value. Another important observation is that at larger time headways (i.e. when time headway exceeds eight seconds), the average travel speed slightly declines. One possible explanation to this decline in mean speed is that vehicles with larger headways typically

include platoon leaders (slower-moving vehicles) as larger gaps in the traffic stream typically exist ahead of those vehicles. One other observation that may be somewhat unexpected is the fact that average travel speed for the time headway range 0-1 second is relatively high at four of the six data sets investigated. This deviation from the general pattern may be explained as follows: vehicles with very short headways (less than one second) are more likely to be associated with aggressive drivers that are either stuck in slow-moving platoons, or traveling at higher speeds following other vehicles while leaving short headways ahead of them. Examination of the speed data has confirmed this main finding. However, it should be mentioned that the number of observations in this headway range is typically small and this was the main reason why this observation was common to only four of the six study sites. Table 3 shows vehicle count and speed standard deviation for each headway range at all study sites. It is clear that the speed standard deviation for headways less than one second is remarkably higher than that for the other headway ranges which is highly consistent with the previous discussion.

Figure 2: Mean Travel Speed versus Time Headway

Ahmed Al-Kaisy and Casey Durbin / Procedia Social and Behavioral Sciences 16 (2011) 329-339 Table 2: Mean Travel Speed and Headway Threshold Relationship

Study Site Mean Travel Speed (y) Mean Headway (x) Correlation Coefficient Best Fit R2

Hwy 287 N-NB 73.17 23.37 0.8359 y = 0.0105x3 - 0.1997x2 + 1.2504x + 72.075 0.9946

Hwy 287 N - SB 63.99 24.83 0.2439 y = 0.0037x3 - 0.0723x2 + 0.4039x + 63.691 0.9974

Hwy 287 S - NB 66.93 30.54 0.9016 y = 0.0041x3 - 0.125x2 + 1.1346x + 65.827 0.9892

Hwy 287 S - SB 70.13 30.39 0.8480 y = 0.004x3 - 0.0946x2+ 0.7081x + 69.487 0.9964

Jackrabbit Ln-NB 62.87 23.17 0.8202 y = 0.0093x3 - 0.176x2 + 1.0886x + 61.928 0.9733

Jackrabbit Ln-SB 58.32 14.19 0.8060 y = 0.0084x3 - 0.1848x2 + 1.2762x + 57.133 0.9839

Table 3: Vehicle Count and Mean Speed Standard Deviation for Various Time Headways

Headway Range

Highway 287 South

Highway 287 North

Jackrabbit Lane

Northbound

Southbound

Northbound

Southbound

Northbound

Southbound

(sec) Count SSD* Count SSD* Count SSD* Count SSD* Count SSD* Count SSD*

0-1 73 9.8 112 9.9 18 11.3 16 13.7 5 18.1 6 20.5

1-2 1087 7.4 1059 6.6 456 8.2 284 5.0 254 6.9 585 6.1

2-3 1329 6.8 908 6.4 394 8.2 328 4.7 314 6.9 736 6.0

3-4 659 6.6 479 6.3 204 7.1 169 4.5 161 7.3 375 6.1

4-5 347 6.2 346 5.9 119 6.7 127 4.4 104 7.5 219 6.4

5-6 239 6.6 227 6.3 91 7.0 103 4.1 61 7.7 141 7.3

6-7 152 6.0 215 5.7 88 6.6 77 4.2 64 5.1 113 7.5

7-8 153 5.1 207 7.2 57 6.6 65 3.4 58 9.8 103 6.2

8-9 134 6.3 153 5.6 56 7.1 82 4.3 53 5.0 74 7.6

9-10 114 5.5 151 6.1 54 6.8 63 4.9 49 5.7 77 7.7

10-11 117 5.5 139 5.8 38 4.5 66 4.0 40 7.8 66 5.5

>11 3556 6.3 4536 6.4 1297 7.8 1292 5.5 965 7.6 996 6.8

Mean Speed Standard Deviation

6.2. Travel Speed versus Platoon Size

In practical terms, vehicular platoons have mostly been defined in terms of time headway. Using this surrogate measure alone, vehicles that travel at relatively high speeds are still part of the platoon as long as the headway of the following vehicle is considered short. On the other hand, platooning on two-lane highways is a major phenomenon that is associated with lack of passing opportunities thus traveling at less than desired speeds being impeded by slow-moving vehicles. A logical hypothesis that is tested in this research is that the type of platoon, i.e. platoons formed by lack of passing opportunities versus those formed by drivers leaving relatively short headways from lead vehicles, is largely a function of platoon size. In other words, it was hypothesized that platoons of larger sizes are mostly associated with slow-moving vehicles restricting speeds of the following vehicles, while small-size platoons (2-3 vehicles) may or may not be related to slow-moving vehicles and the inability to pass.

To test the aforementioned hypothesis, the relationship between platoon size and average speed of platooned vehicles was established as shown in Figure 3. This figure shows mean travel speed of platoons that are equal or

Highway 287 South

Two or m ore Three or m ore Four or more Vehicles in Platoon

JackRabbit Lane

Two or more Three or more Four or more Vehicles ia Platoon

Figure 3: Mean Speed of Platoons of X or More Vehicles

greater in size than a specific number of vehicles using the peak-hour traffic data at each study site. In this analysis, a headway value of 3 seconds is used as a threshold in identifying platooned vehicles. While the previous analysis confirmed that vehicular interaction may exist between successive vehicles with higher headway values, the use of 3-second headway was deemed appropriate to ensure that the majority of vehicles identified are in car-following mode, i.e. considered as platooned vehicles. A trend that is common to all study sites is the larger the platoon size the lower the mean speed. This can be interpreted as follows: the larger the platoon size, the higher the likelihood that a platoon leader is a slow-moving vehicle, and that the following vehicles in the platoon are traveling at lower than their desired speeds.

6.3. Platooned Versus Unimpeded Vehicles

As discussed earlier, any two successive vehicles in the traffic stream separated by a short time headway may indicate one of the following two situations: (1) a lead vehicle impeding the movement of the following vehicle, or (2) two vehicles traveling at or close to their desired speeds with the following vehicle unwilling to pass the lead vehicle. Logically, the first situation is expected to occur more often if the speed of the lead vehicle is below the mean travel speed of all vehicles in the traffic stream, while the second situation is more likely to occur when the lead vehicle is traveling above the mean travel speed. In this research, it is hypothesized that the larger the platoon size, the more the impedance to traffic and the more likely that the first situation would apply to any two successive vehicles separated by a short time headway. By the same token, the smaller the platoon size, the more likely that the second situation would apply to any pair of two successive vehicles separated by a short time headway. To test this hypothesis, speed data was analyzed at all study sites using all speed observations (peak and off-peak intervals). Figure 4 shows mean speed of various size platoons, unimpeded vehicles, and all vehicles for all study sites during the peak hour. The mean speed of unimpeded vehicles (or mean of desired speeds) was estimated by considering those vehicles with headways of eight seconds or more. This was done to ensure that those vehicles are outside the car-following mode (in consistence with results presented in Figure 1).

Upon a quick examination of Figure 4, the following observations cab be made:

1. Mean speed of two-vehicle platoons is generally comparable to the mean speed for all vehicles in the traffic stream. This observation supports the hypothesis that many of two-vehicle platoons are formed by the following vehicle voluntarily leaving a perceived minimum safe headway from the lead vehicle and that the situation does not necessarily involve impedance to traffic.

2. Mean speed of larger size platoons is notably lower than the mean speed of all vehicles which suggests that the amount of impedance to traffic increases with the increase in platoon size.

3. Mean speed of unimpeded vehicles, i.e. vehicles outside of platoons, is notably higher than the mean speed of vehicles that are in platoons. This observation is expected due to the restriction on speeds caused by platoon formation.

4. Mean speed of unimpeded vehicles is noticeably higher that mean speed of overall traffic for all data sets with the exception of HWY 287 North southbound where the two mean speeds are comparable. This is consistent with the explanation provided earlier that the presence of passing lanes upstream of that study site helps to break up platoons and minimize the amount of impedance to traffic.

7. Summary of findings

The research presented in this paper provided valuable insights into the phenomenon of platooning on two-lane two-way highways. A total of six data sets at three study sites in the state of Montana were used in this investigation. Time headway and speed data were investigated to discern the nature of car-following behavior as related to platooning on two-lane highways. The most important findings of this research are:

Figure 4. Speed of Various Size Platoons Relative to Desired Speeds

1. The interaction between successive vehicles on same travel lane mostly diminishes when time headway exceeds a specific value that fell in the range of 5-7 seconds. The implication of this finding is important in identifying vehicles that are outside the influence of platoons and subsequently in estimating desired travel speeds.

2. While short time headway is in most cases an indicator of being impeded by the lead vehicle, this research confirmed that very short headways that are less than one second are more related to aggressive driving, and as such are generally associated with higher speeds.

3. The amount of impedance to traffic is proportional to the size of platoon as evidenced by the relative difference between mean speed of various size platoons and mean speed of unimpeded vehicles.

4. Those and other findings presented in this paper have important implications in modeling traffic operations on two-lane highways and in advancing the current analytical tools including traffic simulation models for two-lane highways.

Acknowledgement: the authors would like to acknowledge the financial support to this research by the Western Transportation Institute of Montana State University.

References

Transportation Research Board. (2000). Highway capacity manual. (4th ed.), TRB, National Research Council, Washington, D.C., 2000.

Luttinen, R. (2001). Percent time-spent-following as performance measure for two-lane highways." In Transportation Research Records: Journal of the Transportation Research Board, No. 1776, TRB, National Research Council, Washington, D.C., pp. 52-59.

Gattis, j., Alguire, M., Townsend, K., and Rao, S. (1997). Rural two-lane passing headways and platooning. In Transportation Research Records: Journal of the Transportation Research Board, No. 1579, TRB, National Research Council, Washington, D.C., pp. 27-34.

Guell, D., and Virkler, M. 1988. Capacity analysis of two-lane highways. In Transportation Research Records: Journal of the Transportation Research Board, No. 1194, TRB, National Research Council, Washington, D.C., pp. 199-205.

Roess, R. (1984). Level of service concept: development, philosophies, and implications. In Transportation Research Record: Journal of the Transportation Research Board, No. 971, TRB, National Research Council, Washington, D.C., pp. 1-6.

Arasan, V., and Kashani, S. (2003). Modeling platoon dispersal pattern of heterogeneous road traffic. In Transportation Research Record: Journal of the Transportation Research Board, No. 1852, TRB, National Research Council, Washington, D.C., pp.175-182.

Guar, A., and Mirchandani. P. (2001). Method for real time recognition of vehicle platoons. In Transportation Research Record: Journal of the Transportation Research Board, No. 1748, TRB, National Research Council, Washington, D.C., pp. 8-17.

Morrall, J. and Werner, A. (1982). Measurement of level of service for two-lane rural highways, Canadian Journal of Civil Engineering, 9, pp.385-398.

Sadeghhosseini, S., and Benekohal, R. (1997). Space headway and safety of platooning highway traffic. Conference Proceedings, Traffic Congestion and Traffic Safety in the 21st Century: Challenges, Innovations, and Opportunities. 8-11 June.

McLean, J. R. (1989). Two-lane highway traffic operations. Gordon and Breach Science Publishers. Melbourne, Australia.