Sage Journals: Discover world-class research

Abstract

This study investigated the operational impact of heavy commercial vehicle (HCV) platooning on urban arterials. HCV platooning is an important application of vehicle-to-vehicle (V2V) technology, with urban arterials facilitating an essential component of HCV movements when picking up and delivering goods. HCV platooning has the potential to reduce fuel consumption and emissions. Moreover, the increasing HCV driver shortage problem can be alleviated if the vehicle following behind a lead vehicle can function without a driver by using autonomous technology enabling Society of Automotive Engineers Level 4 or higher. PTV VISSIM was used to develop a set of micro-simulation models that investigated the impact of traffic signal priority (TSP) and low levels (0%, 5%, and 10%) of HCV platooning. The performance measures include travel time and the number of stops. With the existing traffic control system, HCV platooning increased travel time and increased the number of stops for all vehicles including passenger cars and HCVs. TSP with 5% HCV platooning improved travel time and decreased the number of stops for all vehicles. TSP with 10% HCV platooning, however, only decreased travel time and the number of stops for passenger vehicles. The results suggest that a higher penetration rate of HCV platooning may create significant delays and overwhelm the traffic system even with the assistance of TSP. The findings of this study highlight the potential for TSP to mitigate the impact of HCV platooning on traffic congestion. However, the TSP system may not be a panacea that works for all traffic compositions.

Keywords

freight systems trucking industry research platooning

This study investigates the possible impact of heavy commercial vehicle (HCV) platooning (also known as truck platooning) on urban arterials. A set of PTV VISSIM micro-simulation models have been developed to measure the mobility of HCV platooning with/without traffic signal priority (TSP).

Background

HCVs will prefer to use large highway infrastructure where possible but, in many cases, cannot avoid urban travel for the first-mile and last-mile trip components. In the United States in 2018, HCVs traveled a total of 304,864 million miles, of which 164,321 million miles (54%) were on urban roadways. The urban roadways can be further disaggregated into 66,727 million miles (40.6%) on urban freeways and 97,594 million miles (59.4%) on other urban roadways such as major and minor arterials ( 1 ).

Current vehicle-to-vehicle (V2V) technologies can provide wireless connections that allow two or more HCVs to travel in a platoon ( 2 – 7 ). According to the Society of Automotive Engineers (SAE), HCV platooning is an example of Level 3 to Level 5 automation. Level 3 platooning requires a human driver in every vehicle. Most HCV platooning tests have been conducted at this level to date. At Level 4, a human driver is required in the leading vehicle, but one or more of the following vehicles may be driverless. At Level 5, none of the vehicles in a platoon have a human driver ( 8 – 10 ). Several North American jurisdictions including Alberta, California, Florida, Ontario, and Texas, have permitted Level 3 HCV platoon testing on designated freeway corridors ( 11 , 12 ). The tests have shown that a platoon tightens the gaps between the HCVs and helps to minimize aerodynamic drag, leading to reduced fuel consumption and emissions. At 80 km/h (49.7 mph) on freeways, fuel consumption can be reduced by up to 6% for the lead HCV and by up to 10% for the following HCVs ( 2 – 7 ). For a platoon traveling at 70 km/h (43.5 mph) on a freeway, Al Alam et al. ( 13 ) found a 4.7% reduction in fuel consumption for the lead HCV and a 7.7% reduction for the following HCVs.

HCVs often need to travel on urban arterials when, for example, picking up or delivering goods to warehouses, airports, and intermodal facilities (e.g., railroad yards). Travel on urban arterials involves signalized intersections and posted speed limits which are often lower than 70 km/h (43.5 mph); therefore, reductions in fuel consumption and emissions for HCV platoons on urban arterials are expected to be lower than on freeways. HCV platooning on urban arterial networks may, however, offer benefits, especially when compared with a long combination vehicle (LCV). An LCV is a tractor-trailer combination with two or more trailers driven by a single driver. An LCV is therefore similar to an HCV platoon since both approaches use multiple trailers, but the former can gain fuel savings without the need for autonomous technology. Yet the additional length of an LCV requires more physical space. LCVs require additional horizontal clearance and increased turning radii when compared with standard tractor-trailers, and they therefore encounter more difficulties with many urban intersections and roundabouts ( 14 , 15 ). An HCV platoon can use the same horizontal clearance and turning radius as a single HCV. The increased physical dimensions of an LCV will also require more space at truck parking facilities and LCVs may struggle with existing truck parking availability ( 16 , 17 ).

To ensure safe operations, the requirements for LCV driver licensing are more onerous than those for traditional HCVs. In Ontario, for example, an LCV driver must have at least 5 years of HCV driving experience and more than 100 h of classroom training ( 18 , 19 ). HCV platooning at Level 4 or higher may help to alleviate the shortage of truck drivers without the additional restrictions of LCVs. Costello ( 20 ) reported that the United States will need approximately 160,000 new HCV drivers by 2028, and Butler ( 21 ) reported that Canada will need 34,000 new HCV drivers by 2024. In 2018, FPInnovations tested the technical feasibility of Level 4 HCV platooning on two-lane rural highways connecting logging sites to a nearby port. The platooned HCVs were driven by a single driver in the lead HCV ( 22 ). We found that most of the past studies discussing Level 3 HCV platooning focused on investigating the impact on fuel consumption and emissions. There is a clear research need to investigate the operational impact of Level 4 (or higher) HCV platooning under the existing roadway and traffic environment especially in the urban setting.

Problem Statements

Most HCV platooning studies have been conducted on freeways which are very different from the geometric and traffic environment of arterials and other urban roadways where at-grade intersections make traffic signals and other measures necessary. Numerous studies have investigated special traffic control strategies that include a priority signal phase for a targeted transportation mode such as transit buses, emergency vehicles, or freight transportation ( 23 – 29 ). Changes in travel time are typically used to measure the impact of different traffic control strategies on operational performance. Beak et al. ( 27 ) investigated the impact of signal priority on transit buses on an 8.2 km (5.1 mi) arterial with 13 signalized intersections along Redwood corridor in Salt Lake City, Utah. The study reported that transit signal priority reduced overall network travel time by 15.5 s per vehicle and for transit vehicles travel time was reduced by 8.3 s per vehicle. Kaisar et al. ( 30 ) investigated two independent signal priority strategies, one for transit buses and one for freight vehicles. The study was conducted for traffic passing through nine signalized intersections on a 7.1 km (4.4 mi) corridor in Broward County, Florida. Transit signal priority reduced overall corridor travel times by 18.1 s per vehicle, and freight signal priority reduced travel times by 84.2 s per vehicle. However, transit signal priority reduced travel times by 251.2 s per transit vehicle, and freight signal priority reduced travel times by 175.3 s per freight vehicle. Kaisar et al. ( 30 ) also observed an increase in average travel times for cross-street traffic as a result of the signal priority systems. This increase occurred because the extended green time provided for transit and freight vehicles reduced the green time available to cross-street traffic. The study reported an average travel time increase of 50 s for cross-street traffic movements when freight signal priority was applied.

Research by Tiaprasert et al. ( 31 ) and Wang et al. ( 32 ) investigated TSP strategies for platooned passenger vehicles on urban arterial roadways although these two studies differed in their approach. Tiaprasert et al. ( 31 ) employed an analytical analysis while Wang et al. conducted a micro-simulation. Both studies found that the signal priority strategies reduced the travel time and number of stops at intersections for platooned passenger vehicles. No study has yet investigated the impact of signal priority strategies on HCV platoons, which will behave differently because of differences in vehicle length and acceleration/deceleration rates, and so forth. Automated HCV platoons require further attention as the vehicles need to remain close together for any benefits to result. For such platoons, the green phase at signalized intersections must be able to accommodate more than one HCV. This aspect of HCV platooning, especially Level 4 or higher, has implications for existing and future surface infrastructure and traffic control systems at signalized intersections, as re-establishing an HCV platoon broken by signal interruption is very challenging in urban traffic environments with high traffic volumes. Note that we were unable to find any study that has applied signal priorities for more than one-directional traffic flows at an intersection, such as signal priority applied for both through and left-turning traffic simultaneously. This is not surprising as prioritizing traffic flow in one direction will have a negative trade-off with other directions.

The market penetration rate of platooned vehicles refers to the proportion of vehicles of the same vehicle classification traveling as a convoy and has been identified as a key issue in the operational performance of platooned vehicles ( 33 ). In the case of platooned passenger vehicles, market penetration rates of 5% to 100% have been analyzed, usually by micro-simulation ( 33 – 36 ). Van Arem et al. ( 35 ) showed that a 40% or higher market penetration rate for platooned passenger vehicles on freeways can increase freeway capacity by 3%. Arnaout and Bowling ( 34 ) reported no substantial change in traffic throughput on a freeway if the market penetration rate of platooned passenger vehicles is less than 40%. Lioris et al. ( 37 ) and Smith et al. ( 38 ) investigated the impact of passenger vehicle platooning on arterials. Both studies assumed a 100% market penetration rate and both showed that this condition can double the passenger vehicle throughput. Gordon and Turochy ( 39 ) investigated the operational impact of HCV platooning on a freeway by estimating travel time delay changes on 8.5 km (5.3 mi) of Interstate I-85 in Alabama. They compared simulated travel time delay with baseline travel time delay for traffic flows with no platooning, resulting in 19.1 s per vehicle. The study found that an HCV platooning market penetration rate of 20% reduced travel time delay by 7.7 s (40%) for passenger vehicles and HCVs. A 100% market penetration rate reduced travel time delay by 13.3 s (69%) for passenger vehicles and HCVs. Notice that all existing studies have investigated the operational impact of varying market penetration rates of HCV platooning on freeways and no study has been published to discuss its impact on urban arterials. Knowing, however, that many roadways surrounding major distribution centers and logistic companies are urban arterials with frequent signalized intersections, it is important to assess the impact of market penetration rates for platooned HCVs on urban arterial mobility.

As this study investigates the operational impact of Level 4 HCV platooning, which requires fully functional network-wide vehicle-to-infrastructure (V2I) technology, it is uncertain when, or if, a roadway will exhibit 100% market penetration for platooned vehicles at this level of automation. Although many studies have discussed relatively high market penetration rates for vehicle platooning, we think it is unlikely that Level 4 HCV platooning can be widespread in the foreseeable future as it requires not only further development in vehicle technology but also large capital investments by government to enable seamless wireless communication between numerous traffic control systems installed at many intersections and numerous HCVs traveling on our roadway. Nonetheless, there is a need for investigation into the impact of relatively low percentages of HCV platooning operations on the mobility of urban roadways. A high percentage is not expected in the near future as the HCV platooning technology is first being utilized between HCVs in the same fleet that travel along the same route. This is notable for large establishments requiring multiple trucks to make regular and frequent deliveries between a given origin and destination. In the longer term, a higher penetration rate may occur as the technology sees widespread acceptance and usage. However, this situation is beyond the scope of the paper. The driving attributes of passenger cars and other HCVs will need to be revisited in such a case to accommodate a system-wide change to all vehicles for automated vehicle (AV) adoption.

Study Goal and Objectives

The goal of this study is to investigate the impact of SAE Level 4 HCV platooning on traffic control systems at signalized intersections on arterial roadways. The study has two main objectives:

Compare the impact of Level 4 HCV platooning operations using existing traffic controls, as a base model scenario, with operations using TSP for HCV platoons; and,

Examine the effect of several low market penetration rates (0%, 5%, and 10%) for Level 4 HCV platooning.

The two traffic control scenarios and the three market penetration rates are tested by micro-simulation because Level 4 HCV platooning is still under development so real-world HCV platooning data are not readily available ( 7 , 40 –42). The study uses PTV VISSIM software (Version 2021) as the main simulation tool ( 43 ). Two measures of performance applied to evaluate the results are the changes in travel time and the number of stops for all vehicles, passenger vehicles, and HCVs.

Study Area and Data

The study corridor was a section of Derry Road serving as a major arterial and designated HCV route in Mississauga, Peel Region, Ontario, located northwest of Toronto. Derry Road has some of the highest HCV traffic volumes of all major arterials in Ontario with up to approximately 960 vehicles per hour (vph) per segment ( 44 , 45 ).

Figure 1 shows the 9.2 km (5.7 mi) section of Derry Road selected as the study corridor. The corridor runs along the northern boundary of Toronto Pearson International Airport, which is the largest airport in Canada and one of the largest passenger traffic and HCV traffic attractors in Ontario. The east end of the study corridor provides a connection to Highway 427, a major Ontario freeway with approximately 17,000 Annual Average Daily Truck Traffic (AADTT) carrying large HCV traffic volumes ( 46 ). The corridor has three lanes in each direction and 16 signalized intersections, resulting in 15 segments from Kennedy Road in the west to Goreway Drive in the east. The two green dots in Figure 1 show the ends of the study corridor. The eight red dots show the major intersections connecting Derry Road to other major arterials. The eight blue dots show the remaining minor intersections that are connected, for instance, to nearby small plazas. The study examined westbound traffic only.

Figure 1.

Mid-day traffic volume, heavy commercial vehicle (HCV) percentage, and posted speed limit on the study corridor.

Traffic volumes supplied by Peel Region were collected during the 2016 fall season from September to November. The study relied on Google Maps and Bing Maps satellite images to extract important input parameters needed to develop a VISSIM model. These parameters included, but were not limited to, the number of lanes, posted speed limit, turning radius for each intersection along the corridor, start and end position of tapered lanes used for exclusive left and right turning movements, and the road width. Peel Region supplied additional traffic and operations data contained in Synchro files, including the 2016 hourly turning movement counts (12 months) at all intersections, the percentage of HCVs, and traffic control information including signal control mode (actuated or semi-actuated), signal split, cycle length, signal offset, and coordination information.

The turning movement count data were obtained for the a.m. peak hour (7:15 to 8:15 a.m.), the mid-day hour (1:00 to 2:00 p.m.), and the p.m. peak hour traffic (4:30 to 5:30 p.m.) for each day of the week. Many intersections in the study area were heavily congested during the a.m. and p.m. peak hours. For example, the average traffic volume entering the 16 intersections during the p.m. peak hour was 5,761 vph and the level of service was D or poorer for all major intersections. With such high traffic volumes, an examination of the impact of different traffic control strategies on HCV platooning operation was not feasible as it would not be possible to observe differences at the clearly oversaturated intersections. The mid-day hour was subsequently selected for simulation. The average mid-day traffic volume entering the intersections was 2,997 vph. This is substantially less congested, with 48% lower volume than during the p.m. peak hour. Figure 1 includes details of mid-day traffic volumes. During the mid-day hour, HCVs averaged 16.2% of total traffic at intersections, and the maximum, exhibited at Torbram intersection, was 20.4% of total traffic.

The 24-hour travel time data of the Ministry of Transportation of Ontario (MTO), collected by the Transportation Tomorrow Survey (TTS) ( 47 ), were used to further calibrate the VISSIM model. TTS collects various traffic flow data (traffic volume, travel speed, travel time, etc.) on behalf of government agencies in southern Ontario. The TTS travel time data were collected for Tuesdays to Thursdays in 2016 during the fall season to match relevant seasonal traffic volumes. The Institute of Transportation Engineers (ITE) Handbook ( 48 , 49 ) was used to fill in other relevant input parameters—for example, maximum acceleration rate for passenger vehicles is set as 1.26 m/s² and that for HCVs as 0.3 m/s²; deceleration rates of 3.4 m/s² for passenger vehicles and 1.6 m/s² for HCVs have been assigned.

Model Development

This section discusses three important issues in developing the micro-simulation models for this study: (1) HCV platoon length, (2) input parameters, and (3) model calibration.

HCV Platoon Length

The total length of an HCV platoon can vary widely depending on the number of HCVs in the platoon, the length of each HCV, and the time headway between the platooning vehicles.

Number of HCVs in a Platoon

Studies of HCV platooning have considered platoons of two to five vehicles ( 7 , 50 –52). Ramezani et al. ( 7 ), for instance, investigated platoons of five HCVs on test freeways and assumed that, if necessary, the platooned HCVs could easily leave or join the platoon for the entire journey. The majority of studies of platoons of three or more HCVs have focused on (1) testing Level 3 HCV platooning technology with every HCV containing a human driver, and (2) understanding how HCV platooning affects fuel consumption and emissions ( 2 – 7 ). Urban arterials present HCV platoons with additional challenges such as intersections where, for example, the green phase of a signalized intersection may not provide enough time for a whole platoon to pass through without interruption. This problem naturally increases as the number of HCVs in a platoon increases.

This study considers platoons of two HCVs. This is because (1) it is expected that most imminent HCV platooning operations will be for two HCVs; and (2) platooning with three or more vehicles will not be feasible if two-HCV platoons are found to present major challenges on urban arterials.

Length of HCVs in Platoon

The WB-20 category was selected for all HCVs. A WB-20 vehicle is 22.70 m (74.5 ft) long and is the longest typical single-trailer HCV available in North America ( 14 , 15 ). If a platoon of two WB-20 HCVs can pass a signalized intersection, it is reasonable to expect that any combination of two platooning HCVs will be successful.

Time Headway between Platooning HCVs

Studies of time headways between two consecutive HCVs in a platoon on freeways have used values ranging from 0.6 s to 1.2 s ( 7 , 52 –54). Since slower traveling speeds on arterials mean that platooning vehicles can maintain tighter proximity on urban arterials ( 55 , 56 ), a value of 0.6 s between HCVs was adopted for the time headway between platooning HCVs.

Input Parameters

Car-following and lane-changing models in a micro-simulation traffic model are particularly important to simulate real-world traffic flows as closely as possible. In VISSIM, Wiedemann 74 (W74) is the most popular car-following and lane-changing model to simulate conventional traffic flows on urban arterials, but Wiedemann 99 (W99) is known to be more appropriate for simulating the driving behavior of autonomous vehicles ( 43 , 57 , 58 ). This study used W74 to simulate the behavior of conventional passenger vehicles and HCVs, and W99 to simulate the travel behavior of HCV platoons. Table 1 provides information for the 10 input parameters used in this study. All parameters in Table 1 were input via the PTV VISSIM COM interface and python script ( 43 ).

Table 1.

Simulation Input Parameters

Input parameters	Wiedemann 74			Wiedemann 99
Input parameters	Default	Passenger	HCV	Default	HCV platoon
Maximum look ahead distance (m)	250	250	250	250	300
Maximum look back distance (m)	150	150	150	150	300
Number of interaction objects	4	1	1	2	3
Number of interaction vehicles	99	1	1	99	2
Average standstill distance (m) for a following car	2.0	2.87	3.77	na	3.77
Average standstill distance (m) for a following HCV	2.0	4.07	3.37	na	3.37
Standstill distance (m) for a following HCV platoon	na	na	na	1.5	1.00
Time headway (s)	na	na	na	0.9	0.6
Maximum deceleration for cooperative braking (m/s²)	−3	−3	−1.62	−3	−1.62
Safety distance reduction factor (m)	0.6	0.6	0.6	0.6	1

Note: HCV = heavy commercial vehicle; na = not applicable.

The input parameters are discussed in greater detail below.

The “maximum look ahead distance” and “maximum look back distance” for passenger vehicles and conventional HCVs are the maximum distances a human driver can see ahead or behind when observing the surrounding traffic environment. The distances given are those typically used in previous studies ( 43 ). The value for HCV platooning is identified as the distance over which an autonomous HCV can wirelessly communicate with another autonomous HCV. The 300 m (984.2 ft) distance given for HCV platooning is based on suggestions made by the National Highway Traffic Safety Administration (NHTSA) ( 59 ).

The “number of interaction objects” refers to a vehicle’s interaction (i.e., communications) frequency with other objects such as traffic controls and nearby vehicles. The “number of interaction vehicles” indicates a subset containing only the number of interactions with other vehicles. As conventional passenger vehicles and HCVs are controlled by humans, this study set the number of interaction objects and number of interaction vehicles for both passenger cars and conventional HCVs to a value of 1 as suggested by PTV ( 43 ). This means that a human controlled vehicle interacts (communicates) with other objects one at a time. The number of interaction objects and number of interaction vehicles for HCV platooning are higher than 1 because an HCV in a platoon needs to communicate simultaneously and automatically with other HCVs in the same platoon and with upcoming traffic controls ( 60 ).

The “average standstill distance for a following car” and the “average standstill distance for a following HCV” indicate the minimum distance required to avoid a collision between a lead vehicle and following vehicle. The model uses 2.87 m (9.42 ft) for the average standstill distance between two passenger vehicles and 3.37 m (11.06 ft) for the average standstill distance between two consecutive HCVs as suggested by Lu et al. ( 61 ).

The “standstill distance for a following HCV platoon” indicates the minimum distance required to avoid a collision between two HCVs in the same platoon. This study used 1.00 m (3.28 ft) as suggested by Deng ( 62 ) and Deng and Boughout ( 63 ).

The “time headway” indicates the distance from the front of the leading vehicle or object to the front of the following vehicle and is additional to the standstill distance. A value of 0.6 s was defined for platooning HCVs.

The “maximum deceleration rate for cooperative braking” indicates the deceleration rate required for a target vehicle to allow another vehicle to change lane and enter the target vehicle’s lane. This study used Harwood et al.’s ( 49 ) value of −1.62 m/s² (−5.31 ft/s²) for conventional HCVs and platooned HCVs.

The “safety distance reduction factor” is a factor applied to the lane-changing model to reduce the minimum safety distance required between the lead and following vehicles when the lead vehicle initiates a lane change. A value of 0.6 indicates that when a following vehicle changes to a target lane, the following vehicle can accept a 40% reduction in the safety distance with the lead vehicle when completing the lane change maneuver. Ahmed et al. ( 64 ), however, proposed using a value of 1 as the value for HCVs in a platoon to reflect the conservative behavior of HCV platoons. A value of 1 means that platooned HCVs make no compromise with safety distance and maintain full minimum safety distance when changing lanes.

Model Calibration

A base model representing the do-nothing scenario has been designed to replicate real-world traffic flow conditions and variability. A micro-simulation model is stochastic in nature as the values of many parameters such as acceleration/deceleration rate and vehicle speed are randomly generated from an assumed probabilistic distribution for each parameter. As a result, each simulation generates somewhat different results and multiple simulation runs are required to obtain reliable results. This study followed the Wisconsin Department of Transportation’s (WisDOT’s) micro-simulation guidelines closely. Using these guidelines, 30 simulation runs were produced for the base scenario. The model produced traffic volumes that simulated an hour of real-world traffic volumes at the 95% confidence level. The two main measures of effectiveness (MOEs) for calibration were traffic volume and travel time. As suggested in the WisDOT guidelines, we used the GEH (Geoffrey E. Harver) statistic to validate traffic volume and correlation (ρ-value) to validate travel time ( 65 ).

Traffic Volume

The GEH statistic is used to model goodness-of-fit ( 65 , 66 ). In this study, the GEH statistic was used to assess the similarity between observed and simulated mid-day traffic. A GEH of less than 5 $(\sqrt{vehicle / hour})$ is usually considered a good fit ( 67 , 68 ). Equation 1 shows the formula:

GEH = \sqrt{\frac{2 {(m - c)}^{2}}{m + c}}

(1)

where

$m$ = simulated mid-day traffic volume for an hour; and

$c$ = observed mid-day traffic volume for an hour.

Figure 2 compares the base model results of observed traffic volumes (green bars) and simulated traffic volumes (white bars). The data include all 16 intersections on the study corridor. For direct comparison, the results include the percentage difference between the observed and simulated traffic volumes and the resulting GEH value. At most intersections, the base model slightly over-estimated the traffic volume, but the differences were small, ranging from 1.06% (Airport Road) to 4.32% (Southbound Highway 410 intersection). The GEH is well under 5 for all 16 intersections, with the highest GEH value of 1.86 corresponding to the Southbound (SB) Highway 410 intersection. These results suggest that the base model adequately simulated real-world traffic volumes.

Figure 2.

Calibration results for base scenario traffic volumes.

Travel Time

A second calibration was performed using average travel time as another MOE, which includes traffic control delay time. Mid-day travel time data were used for a typical weekday average travel time between 1:00 and 2:00 p.m. that measured in seconds for each hour, but data were only available for nine study corridor segments for through traffic movement only. These nine segments included seven sections between the eight major intersections shown as red dots in Figure 1. Two additional segments with sufficient data included the eastern edge of the study corridor to Kennedy Road, and Goreway Drive to the western edge.

Figure 3 compares the travel times observed mid-day with the base model for hourly simulated travel times on the nine segments. The simulated travel times were computed by averaging the results of 30 simulation runs. Each small circle in Figure 3 represents the observed travel times (see x-axis) and the simulated average travel times (see y-axis) for one of nine segments of the corridor. In Figure 3, the green circle, for example, represents 163 s for the observed travel time and 160 s for the simulated average travel time for the segment from Kennedy Road to Tomken Road. The orange circle shows the shortest travel time (50 s for the observed travel time and 49 s for the simulated average travel time for the segment from Goreway Drive to Rexwood Road). The observed and simulated travel times show a very strong positive correlation (ρ = 0.983) and indicate an accurate representation of the real-world travel times.

Figure 3.

Calibration results for base scenario travel times.

Development of Test Scenarios

A set of scenarios were developed to investigate the operational impact of Level 4 HCV platooning on urban arterials with existing traffic controls and with the addition of TSP for HCV platoons. These scenarios included multiple tests to analyze 0%, 5%, and 10% market penetration rates.

The TSP investigated in this study was designed to provide an extended green phase allowing all platooned HCVs to pass through an intersection together. This TSP approach was installed at each of the eight major intersections on the study corridor (the eight red dots in Figure 1) to help through traffic only. As a result, the TSP in this study is not intended to prioritize left-turn or right-turn vehicle movements. Eight minor intersections on the study corridor (the eight blue dots in Figure 1) were assumed to operate using existing signal phasing. The existing signal control timing had longer phases for traffic on Derry Road to accommodate the high levels of traffic along the road when compared with the intersecting minor roads. As a result, this study developed a total of five VISSIM models: the base model and four alternative models. NP refers to existing traffic controls with no TSP, TP refers to TSP conditions, and the number after NP or TP refers to the HCV platooning percentage (5% or 10%). The five models were:

Base model (NP0), representing a do-nothing scenario with no TSP and 0% HCV platooning;

Alternative model 1 (NP5), simulating existing traffic controls with no TSP and the truck volume adjusted to include 5% HCV platooning;

Alternative model 2 (NP10), simulating existing traffic controls with no TSP and the truck volume adjusted to include 10% HCV platooning;

Alternative model 3 (TP5), simulating a TSP system with the truck volume adjusted to include 5% HCV platooning; and

Alternative model 4 (TP10), simulating a TSP system with the truck volume adjusted to include 10% HCV platooning.

TP5 and TP10 used the ring barrier controller and vehicle detection function (simulated loop detectors) embedded in VISSIM. In a micro-simulation model involving a TSP system, the location of the loop detectors before each intersection is important as it determines the length of the extended green. In this study, the location of the loop detectors determines the length of the extended green allowed for a platoon of HCVs to pass through the intersection uninterrupted. According to Kaisar et al. ( 30 ), most simulations use the minimum stopping sight distance (MSSD) defined by the American Association of State Highway and Transportation Ofﬁcials (AASHTO) ( 15 ) to determine the location of the loop detectors. We used the posted speed limit for each segment to estimate the MSSD of each segment and to locate the appropriate position for the loop detectors. VISSIM refers to the simulated loop detector as the check-in detector.

If we use the segment approaching Dixie Road (see Figure 1) as an example, the posted speed limit is 70 km/h (43.5 mph) which gives an estimated MSSD of 105 m (344.5 ft). The location of the simulated loop detectors is therefore set as 105 m (344.5 ft) before the Dixie Road intersection stop-line. In this instance, the ring barrier controller extends the green time by 9 s, with 6 s covering the 105 m (344.5 ft) travel time from the loop detectors to the stop bar and an additional 3 s of slack time. The slack time accounts for variability in the arrival times of the approaching HCVs in the platoon.

Analysis and Results

The study conducted 150 simulations in total with 30 runs completed for each of the five models NP0, NP5, NP10, TP5, and TP10. The two measures of performance consisted of changes in average travel time as a primary measure and the number of stops at traffic lights as a secondary measure ( 69 ). The measures were used to explore the impact of HCV platooning on three vehicle categories: (1) all vehicles, (2) passenger vehicles, and (3) conventional or platooned HCVs for traveling through 16 signalized intersections.

Average Travel Time MOE

Figure 4 is a set of boxplots showing the average travel time of the three vehicle categories. A typical boxplot shows median, first and third quartile values from the median, maximum and minimum values, but Figure 4 shows the mean value $(μ)$ (the middle thick horizontal line in the middle of the box), the mean value ± one standard deviation ( $μ \pm σ)$ (the lower and upper limit of each box), and the mean value ± two standard deviations ( $μ \pm 2 σ)$ (the minimum and maximum value of each vertical line). $μ \pm 2 σ$ can be regarded as the estimated 95% confidence interval (CI). For example, the base model’s mean average travel time (based on 30 simulations) for individual passenger vehicles was slightly less than 14.0 min with a CI ranging from 13.32 min to 14.64 min.

Figure 4.

Average travel times estimated by the five models: (a) all vehicles, (b) passenger vehicles, and (c) HCVs.

The information shown in Figure 4 was used as to conduct a set of two-sample Welch’s t-tests ( 70 ), which are also known as unequal variance t-tests. This provided a statistical basis for comparing the average travel times of the base model with the average travel times of the four alternative models. The null hypothesis was that the average travel time for each alternative model was equal to the base model’s travel time. The t-statistic and degrees of freedom were calculated using Equations 2 and 3:

t = \frac{μ_{1} - μ_{2}}{\sqrt{\frac{σ_{1}^{2}}{n_{1}} + \frac{σ_{2}^{2}}{n_{2}}}}

(2)

df = \frac{{(\frac{σ_{1}^{2}}{n_{1}} + \frac{σ_{2}^{2}}{n_{2}})}^{2}}{\frac{σ_{1}^{4}}{n_{1}^{2} (n_{1} - 1)} + \frac{σ_{2}^{4}}{n_{2}^{2} (n_{2} - 1)}}

(3)

where

$t$ = t-statistic;

$μ_{1}$ = base model’s average travel time for passenger vehicles, HCVs (including platooned HCVs), and all vehicles;

$μ_{2}$ = each alternative model’s average travel time for passenger vehicles, HCVs (including platooned HCVs), and all vehicles;

$σ_{1}$ = base model’s standard deviation for travel time for passenger vehicles, HCVs (including platooned HCVs), and all vehicles;

$σ_{2}$ = each alternative model’s standard deviation for travel time for passenger vehicles, HCVs (including platooned HCVs), and all vehicles;

$n_{1}$ = number of observations (30) in the base model; and

$n_{2}$ = number of observations (30) in each alternative model.

$df$ = degrees of freedom.

Table 2 summarizes the results of the t-tests. A discussion of the results, from Figure 4 and Table 2, is presented below.

Table 2.

T-Test Results for Average Travel Time (min)

Vehicle category	VISSIM model	Mean	SD	df	t-statistic	p-value
All vehicles	NP0	14.23	0.36	na	na	na
	NP5	14.63	0.37	57.88	−4.19	9.60E-05***
	NP10	14.78	0.40	57.19	−5.58	6.80E-07***
	TP5	13.75	0.30	56.52	5.67	5.01E-07***
	TP10	14.03	0.26	53.12	2.56	0.013**
Passenger vehicles	NP0	13.98	0.33	na	na	na
	NP5	14.15	0.29	56.99	−2.14	0.04**
	NP10	14.26	0.29	57.09	−3.55	7.71E-04***
	TP5	13.31	0.15	41.27	10.06	1.16E-12***
	TP10	13.45	0.18	45.41	7.73	7.99E-10***
HCVs	NP0	14.49	0.40	na	na	na
	NP5	15.11	0.60	50.53	−4.70	2.04E-05***
	NP10	15.30	0.60	50.82	−6.20	1.00E-07***
	TP5	14.18	0.58	51.69	2.39	0.02**
	TP10	14.60	0.43	57.80	−1.08	0.29

Note: SD = standard deviation; df = degrees of freedom; na = not applicable; HCV = heavy commercial vehicle. Significance level: ***<0.01; **<0.05; *<0.1.

All Vehicles

For all vehicles, the boxplots in Figure 4a and tabular results in Table 2 show that NP0’s average travel time was 14.23 min. NP5 (existing traffic controls with 5% HCV platooning) increased travel time by 24 s (≈ (14.63 min −14.23 min) × 60 s) per vehicle. NP10 (existing traffic controls with 10% HCV platooning) increased travel time by 33 s per vehicle. The t-tests showed that the increased travel times for all vehicles were statistically significant at the 99.9% confidence level. These results indicate that a small percentage of HCV platooning without improved traffic control systems created additional delays on the study corridor.

TP5 (TSP with 5% HCV platooning) reduced travel time for all vehicles by 29 s per vehicle, and TP10 (TSP with 10% HCV platooning) reduced travel time for all vehicles by 12 s per vehicle. The t-tests showed that the decreased travel times for all vehicles were statistically significant at the 99.9% confidence level for TP5 and at the 95% confidence level for TP10. These results suggest that a TSP strategy with 5% to 10% of HCV platooning can improve the travel times for all vehicles on the corridor.

Passenger Vehicles

Figure 4b shows the pattern for passenger vehicle travel times. With the existing traffic control system, 5% HCV platooning (NP5) increased passenger vehicle travel times by 10 s per vehicle, and 10% HCV platooning (NP10) increased passenger vehicle travel times by 17 s per vehicle. The t-tests showed that the increased passenger vehicle travel times were statistically significant at the 95% confidence level for 5% HCV platooning and at the 99.9% confidence level for 10% HCV platooning. TP5 reduced passenger vehicle travel times by 40 s per vehicle, and TP10 reduced passenger vehicle travel times by 32 s per vehicle. Saving 40 s of travel time could allow a passenger vehicle to travel an extra 670 m (60 km/h) to 780 m (70 km/h). The t-tests showed that the TP5 and TP10 results were statistically significant at the 99.9% confidence level, and suggest that an extended green phase targeting platooned HCVs could also help passenger vehicles to pass through major intersections with reduced delay.

HCVs

Figure 4c shows that the pattern for HCVs was similar to the patterns for all vehicles and passenger vehicles for three of the alternative models: compared with NP0 for HCVs, NP5’s travel time increased by 37 s per vehicle, NP10’s travel time increased by 49 s per vehicle, and TP5’s travel time decreased by 19 s per vehicle. TP10, however, increased rather than decreased HCV travel time. The increase was 7 s. TP10 did not improve HCV travel time and this model failed to reject the null hypothesis. This result suggests that a 10% or higher rate of HCV platooning may create significant delays, especially for HCVs, even with TSP installed at all the study corridor’s major intersections. Notice, however, that the TSP strategy is able to effectively offset the negative impacts on traffic congestion without traffic control improvement with the same percentage (10%) of HCV platoon shown in NP10.

The table given in the appendix shows the directional movements at the major intersections along the study corridor, and summarizes the average travel time results of all vehicles for the NP0, NP5, NP10, TP5, and TP10 models. The table also shows the percentage difference from the base model (NP0).

The results in the appendix show increased travel times for cross-street traffic for 5% and 10% HCV platooning. The average travel times for 10% HCV platooning were higher than for 5% HCV platooning. This result is consistent with the results shown in Table 2 and Figure 4. The results are similar for the TP5 and TP10 signal priority scenarios, that is, average travel times increase with 5% HCV platooning and especially with 10% HCV platooning.

Average travel times for each of the eight signalized cross-street intersections with signal priority decreased by 5 s per vehicle from NP0 to TP5. In the case of cross-streets, average travel times increased by 0.6 s per vehicle from NP0 to TP10, a negligible difference. For comparison, Kaisar et al. ( 30 ) observed a 50 s average travel time increase for cross-street traffic for non-platoon HCV signal priority. We hypothesize that the larger travel times in their study occurred because of higher traffic volumes.

Number of Stops MOE

Figure 5 is a set of boxplots showing changes in the number of stops for the three different vehicle categories. The number of stops directly affects fuel consumption and greenhouse gas (GHG) emission ( 69 , 71 ). With similar formatting to the previous boxplots, Figure 5 shows the average, ±1 standard deviation, and ±2 standard deviations for the number of stops in each of the five models. Table 3 shows the results of the two-sample Welch’s t-tests. The null hypothesis was that the number of stops of each alternative model would be equal to the number of stops of the base model.

Table 3.

T-Tests Results for Number of Stops

Vehicle category	VISSIM model	Mean	SD	df	t-statistic	p-value
All vehicles	NP0	5.18	0.36	na	na	na
	NP5	5.44	0.32	57.24	−3.00	4.01E-03**
	NP10	5.51	0.34	57.87	−3.66	5.47E-04***
	TP5	4.85	0.21	48.76	4.10	1.58E-04***
	TP10	5.20	0.33	57.62	−0.26	0.8
Passenger vehicles	NP0	5.21	0.33	na	na	na
	NP5	5.31	0.26	54.74	−1.31	0.2
	NP10	5.36	0.26	55.12	−1.99	0.05*
	TP5	4.85	0.23	51.05	4.90	1.02E-05***
	TP10	5.07	0.19	46.27	2.00	0.05*
HCVs	NP0	5.16	0.39	na	na	na
	NP5	5.58	0.52	54.02	−3.55	7.96E-04***
	NP10	5.66	0.47	56.34	−4.53	3.09E-05***
	TP5	4.86	0.42	51.68	2.69	9.59E-03**
	TP10	5.34	0.57	51.52	−1.47	0.15

Note: SD = standard deviation; df = degrees of freedom; na = not applicable; HCV = heavy commercial vehicle. Significance level: ***<0.01; **<0.05; *<0.1.

Figure 5.

Number of stops estimated by the five models: (a) all vehicles, (b) passenger vehicles, and (c) HCVs.

The results shown in Figure 5 and Table 3 are discussed below.

All Vehicles

Compared with NP0, the number of stops for both NP5 and NP10 significantly increased. NP10, for instance, required all vehicles to stop an average of 0.34 (= 5.51 − 5.18) additional stops when traveling along the study corridor. As the mid-day hour’s average traffic volume on the corridor is close to 3,000 vph, NP10 would introduce approximately 990 additional stops per hour along the study corridor by all vehicles compared with NP0. TP10 also significantly increased the number of stops for all vehicles, but the change was not statistically significant. Only TP5 decreased the number of stops for all vehicles. The decrease was statistically significant at the 99.9% confidence level.

Passenger Vehicles

NP5, NP10, and TP10 increased the number of stops for passenger vehicles. The increases were statistically significant for NP10 and TP10 (both with 90% confidence level). Only TP5 decreased the number of stops for passenger vehicles. The decrease was statistically significant at the 99.9% confidence level.

HCVs

The HCV results followed a similar pattern, with NP5, NP10, and TP10 increasing the number of stops for HCVs. The NP5 and NP10 increases were statistically significant at the 99.9% confidence level. Only TP5 decreased the number of stops for HCVs. The decrease was statistically significant at the 95% confidence level.

Analysis Results Summary

TP5 produced different results from NP5, NP10, and TP10. It was the only alternative model to show a statistically significant decrease in both the travel time and the number of stops for all three vehicle categories. This result suggests that TSP with 5% HCV platooning may help to reduce travel times, fuel consumption, and GHG emissions. TSP with 10% HCV platooning (TP10), however, increased the number of stops, though not as much as NP5 and NP10, suggesting that the extended green phase could not handle a 10% (or higher) market penetration rate for platooned HCVs when compared with NP0. Not all the platooned HCVs will be able to cross the signalized intersection within the 9 s extended green time of the cycle length. Therefore, as the rate of HCV platooning increases there will be an increased likelihood of some vehicles not crossing the intersection in time.

In summary, this study attempted to observe the impact of HCV platooning at low levels (5% and 10%) on urban arterial roadways as a step toward understanding the consequences of introducing HCV platooning operation on urban arterials in the near future. HCV platooning penetration is expected to increase slowly as V2X (vehicle-to-everything) technology progresses. It is, however, reasonable to expect that because of unforeseen technological shifts, the accuracy of current hypothetical scenarios will decrease for predictions made further into the future.

Conclusion

The goal of this study was to investigate the impact on mobility of SAE Level 4 HCV platooning on traffic control systems at signalized intersections on arterial roadways. Level 4 HCV platooning requires a human driver in the lead vehicle, but one or more closely following HCVs may be driverless. The study is the first to investigate the operation of Level 4 HCV platooning on arterials with signalized intersections. It was conducted using VISSIM micro-simulations comparing the base model (0% HCV platooning) with four alternatives. These alternative models included: existing traffic controls with 5% HCV platooning, existing traffic controls with 10% HCV platooning, a TSP system with 5% HCV platooning, and a TSP system with 10% HCV platooning. Two measures of performance were used, namely the average travel time and the number of stops. These measures were assessed with respect to all vehicles, passenger vehicles, and HCVs. The study corridor was a known HCV-heavy corridor in Peel Region, Ontario, Canada. The main findings were:

As Level 4 HCV platooning has communication capability with the traffic controllers and runs with a tight headway gap, initially we expected that with the increase of HCV platooning the operational performance might improve. However, under the existing traffic control environment, the introduction of HCV platooning on the study corridor would lead to a significant deterioration in the corridor’s operational performance especially for HCVs. The findings were observed for low levels of HCV platooning (5% and 10%) yet worsening conditions can be expected if the rate increases further. The implication is that our current surface infrastructure is not yet ready for the introduction of HCV platooning operations on urban arterials; and,

If signalized intersections can be improved by providing all major intersections with TSP for HCV platooning, it may be possible to allow up to approximately 5% HCV platooning on the tested roadway corridor of Derry Road. At this level of HCV platooning, the average travel times and the number of stops were similar to those of the base (do-nothing) model. Before conducting the micro-simulation analyses, we expected TSP would be helpful for both 5% and 10% HCV platooning for improving the operational performances. This result also indicates that strategies such as TSP can mitigate the negative implications of 5% HCV platooning.

A cost–benefit analysis can be applied to evaluate travel time changes from TSP for HCV platooning. The value of time for passenger vehicle occupants and the value of time for HCV drivers are not the same, since goods shipped by HCVs add extra costs from travel delays. For example, Kaisar et al. ( 30 ) estimated value of time for an HCV and passenger vehicles to be $80 per hour and $15 per hour, respectively. Based on that, travel time savings in our study have been converted to dollar value savings. The highest benefit is estimated for the 5% HCV platooning under TSP condition with all vehicles saving an average of 3.4%. Similarly, passenger vehicles receive 4.8% savings and HCVs receive 2.1% savings for 5% HCV platooning operation under TSP. However, this benefit is only 1.4% savings for all vehicles for operation of 10% HCV platooning under TSP condition.

The analysis made by this study suggests that HCV platooning under existing traffic control conditions will reduce mobility on urban arterial roadways. An HCV platoon will be more likely than a single HCV to have to stop at an intersection because of the additional length requiring more green time. The average travel time and the number of stops will increase with a rise in the HCV platooning penetration rate. The application of TSP at a 5% HCV platooning rate will improve the mobility and operational performance of the selected urban arterial corridor, but this improvement will not occur if the penetration rate exceeds 5%. An alternative technique, such as a dedicated HCV lane, may be better suited to this situation. However, the cost of construction of a dedicated lane may be higher than the value of the time saved.

Existing passenger vehicles and HCVs are still mostly human-driven and do not have interaction capabilities with HCV platooning. Increased V2X communication will improve mobility by reducing the average travel time and the number of stops. This technology may be a cost-effective method for offsetting the negative operational impact of HCV platooning demonstrated in this paper.

Future research could consider the limitations of this study and improve our understanding of the operational impact of HCV platooning on urban arterials. For example:

The TSP system here was assumed to have perfect detection of two consecutive HCVs’ movements as a platoon. Existing TSP systems are not currently designed to detect platooned HCVs.

This study considered TSP for platooned HCVs for through traffic movements only. We note, however, that a future study may benefit from the evaluation of TSP for platooned HCVs on other turning movements (e.g., left-turning HCV platooning). The investigation of TSP for other turning movements will generate valuable inputs for more rigorous analyses that can assist in selecting the most suitable intersections as well as the most suitable directional traffic flows for maximum benefit.

Future studies should conduct a detailed investigation into the different levels of traffic volume and the level of service of the intersection to accommodate HCV platooning operations. The benefits of TSP will be diminished if the intersections are close to or exceeding the capacity of the infrastructure.

Future studies should conduct rigorous benefit and cost analyses to justify a decision to install TSP for HCV platooning at targeted signalized intersections. The analyses need to include the costs associated with the installation, maintenance, and operation of TSP at numerous signalized intersections along the HCV platooning route.

In the current study of urban arterial road conditions, we have considered the relatively near future in which passenger vehicles are expected to remain human-driven and the maximum HCV platoon rate is not expected to exceed 10%. We think that it is not realistic to evaluate the distant future and scenarios such as 50% HCV platooning until V2X technology is adapted for vehicle classifications (e.g., passenger vehicles). We expect that when such technology is available, future studies will consider a higher percentage of HCV platooning.

Future studies should perform pairwise comparison tests to quantify whether the differences between the considered models, such as NP5 and NP10, or NP5 and TP5, or NP0 and TP10, and so forth, are statistically significant.

Future studies need to provide a framework for selecting the urban arterials where TSP for HCV platooning can deliver the greatest benefits.

The results of this study will help transportation engineers and decision makers to understand the mobility challenges associated with HCV platooning on urban arterials.

Supplemental Material

sj-docx-1-trr-10.1177_03611981221127287 – Supplemental material for Operational Impact of the Through-Traffic Signal Prioritization for Heavy Commercial Vehicle Platooning on Urban Arterials

Supplemental material, sj-docx-1-trr-10.1177_03611981221127287 for Operational Impact of the Through-Traffic Signal Prioritization for Heavy Commercial Vehicle Platooning on Urban Arterials by Tanvir Chowdhury, Peter Y. Park and Kevin Gingerich in Transportation Research Record

Footnotes

Author Contributions

The authors confirm the following contributions to this paper: study conception and design: Tanvir Chowdhury, Peter Y. Park; data collection: Tanvir Chowdhury; analysis and interpretation of results: Tanvir Chowdhury, Peter Y. Park; and draft manuscript preparation: Tanvir Chowdhury, Peter Y. Park and Kevin Gingerich. All authors reviewed the results and approved the final version of the manuscript.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors thank the Region of Peel, Ontario, Canada and the Smart Freight Centre for their financial and other support for this study. We also thank the Natural Sciences and Engineering Research Council (NSERC) for financial contributions.

ORCID iDs

Tanvir Chowdhury

Peter Y. Park

Supplemental Material

Supplemental material for this article is available online.

References

U.S. DOT. Truck Profile. Bureau of Transportation Statistics, 2020.

Browand

McArthur

Radovich

Fuel Saving Achieved in the Field Test of Two Tandem Trucks. Technical Report UCB-ITSPRR-2004-20. California PATH Program, Institute of Transportation Studies, University of California, Berkeley, 2004.

Alam

Besselink

Turri

Martensson

Johansson

K. H.

Heavy-Duty Vehicle Platooning for Sustainable Freight Transportation: A Cooperative Method to Enhance Safety and Efficiency. IEEE Control System Magazine, Vol. 35, No. 6, 2015, pp. 34–56.

Patten

McAuliffe

Mayda

Tanguay

Review of Aerodynamic Drag Reduction Devices for Heavy Trucks and Buses. NRC Technical Report CSTT-HVC-TR-205. National Research Council of Canada, Ottawa, ON, 2012, p. 3.

Zabat

Stabile

Farascaroli

Browand

The Aerodynamic Performance of Platoons: A Final Report. California PATH Program, Institute of Transportation Studies, University of California, Berkeley, 1995.

Tavasszy

L. A.

Janssen

The Value Case for Truck Platooning. Working Paper. ITS European Congress, Glasgow, January, 2017.

Ramezani

Shladover

S. E.

Chou

Microsimulation Framework to Explore Impact of Truck Platooning on Traffic Operation and Energy Consumption: Development and Case Study. University of California, Berkeley, 2018.

U.S. DOT. Automated Vehicles: Truck Platooning: Its Benefits, Costs, and Lessons Learned: 2018 Update Report. Intelligent Transportation Systems Joint Program Office, U.S. Department of Transportation, Washington, D.C., 2018.

Bishop

New Moves, New Markets for Truck Platooning Revealed at AVS2020. Forbes: Transportation, 2020.

10.

SAE International Standard. Automated Driving: Levels of Driving Automation. SAE J3016. Society of Automotive Engineers International, Warrendale, PA, 2016.

11.

Kuhn

Lukuc

Poorsartep

Wagner

Balke

K. N.

Middleton

Songchitruksa

Wood

Moran

Commercial Truck Platooning Demonstration in Texas—Level 2 Automation. Texas Department of Transportation, Research and Technology Implementation Office, Austin, 2017.

12.

CCMTA. Automated Vehicles in Canada. White Paper. Canadian Council of Motor Transport Administrators, Ottawa, ON, 2016.

13.

Al Alam

Gattami

Johansson

K. H.

An Experimental Study on the Fuel Reduction Potential of Heavy Duty Vehicle Platooning. Proc., 13th International IEEE Conference on Intelligent Transportation Systems, Funchal, Portugal, IEEE, New York, 2010, pp. 306–311.

14.

TAC. Geometric Design Guide for Canadian Roads. Transportation Association of Canada, Ottawa, ON, 2017.

15.

AASHTO. A Policy on Geometric Design of Highways and Streets, 7th ed. American Association of State Highway and Transportation Ofﬁcials, Washington, DC, 2018.

16.

Park

Why Waste Time on Platooning When LCVS Are More Efficient?

Heavy Duty Trucking, 2019.

17.

Nevland

E. A.

Gingerich

Park

P. Y.

A Data-Driven Systematic Approach for Identifying and Classifying Long-Haul Truck Parking Locations. Transportation Policy, Vol. 96, 2020, pp. 48–59.

18.

MTO. Long Combination Vehicle (LCV) Program Conditions. Ontario Ministry of Transportation, Toronto, 2021.

19.

Smith

G. J.

Canada’s Truck Drivers to Require 103.5 Hours of Training. trucknews.com, 2020.

20.

Costello

Truck Driver Shortage Analysis. The American Trucking Associations, Arlington, VA, 2019.

21.

Butler

A “Demographic Tsunami” Is About to Make Canada’s Trucker Shortage Even Worse. CBC News, 2019.

22.

Proust

Smith

Pierce

Ward

Bevly

Evaluation of Platooning on Resource Roads. Proc., 2019 TAC-ITS Canada Joint Conference, Halifax, 2019.

23.

Balke

K. N.

Dudek

C. L.

Urbanik

Development and Evaluation of Intelligent Bus Priority Concept. Transportation Research Record: Journal of the Transportation Research Board, 2000. 1727: 12–19.

24.

Kuang

Real-Time Traffic Signal Intelligent Control with Transit-Priority. Journal of Software, Vol. 7, No. 8, 2012, pp. 1738–1743.

25.

Guler

S. I.

Menendez

Analytical Formulation and Empirical Evaluation of Pre-Signals for Bus Priority. Transportation Research Part B: Methodological, Vol. 64, 2014, pp. 41–53.

26.

Rodriguez

Danaher

A. R.

Operational Comparison of Transit Signal Priority Strategies. Transportation Research Record: Journal of the Transportation Research Board, 2014. 2418: 84–91.

27.

Beak

Zamanipour

Head

K. L.

Leonard

Peer-to-Peer Priority Signal Control Strategy in a Connected Vehicle Environment. Transportation Research Record: Journal of the Transportation Research Board, 2018. 2672: 15–26.

28.

Islam

H. L.

Hoang

N. H.

Cricenti

A Linear Bus Rapid Transit with Transit Signal Priority Formulation. Transportation Research Part E: Logistics and Transportation Review, Vol. 114, 2018, pp. 163–184.

29.

Ikiriko

Chimba

Mbuya

Impact of Transit Signal Priority on Isolated Intersection Operations: A Microscopic Simulation Evaluation. Presented at 98th Annual Meeting of the Transportation Research Board, Washington, D.C., 2019.

30.

Kaisar

E. I.

Hadi

Ardalan

Iqbal

M. S.

Evaluation of Freight and Transit Signal Priority Strategies for Improving Transportation Operations in Urban Corridors. Florida Atlantic University, Boca Raton, 2020.

31.

Tiaprasert

Zhang

Platoon Recognition Using Connected Vehicle Technology. Journal of Intelligent Transportation System, Vol. 23, No. 1, 2019, pp. 12–27.

32.

Wang

Jiang

Xiao

Zhao

A Joint Control Model for Connected Vehicle Platoon and Arterial Signal Coordination. Journal of Intelligent Transportation System, Vol. 24, No. 1, 2020, pp. 81–92.

33.

Bujanovic

Lochrane

Capacity Predictions and Capacity Passenger Car Equivalents of Platooning Vehicles on Basic Segments. Journal of Transportation Engineering, Part A: Systems, Vol. 144, No. 10, 2018, p. 4018063.

34.

Arnaout

Bowling

Towards Reducing Traffic Congestion Using Cooperative Adaptive Cruise Control on a Freeway with a Ramp. Journal of Industrial Engineering and Management, Vol. 4, No. 4, 2011, pp. 699–717.

35.

Van Arem

Van-Driel

C. J. G.

Visser

The Impact of Cooperative Adaptive Cruise Control on Traffic-Flow Characteristics. IEEE Transactions on Intelligent Transportation Systems, Vol. 7, No. 4, 2006, pp. 429–436.

36.

Zhao

Sun

Simulation Framework for Vehicle Platooning and Car-Following Behaviors under Connected-Vehicle Environment. Procedia-Social Behaviour Science, Vol. 96, 2013, pp. 914–924.

37.

Lioris

Pedarsani

Tascikaraoglu

F. Y.

Varaiya

Platoons of Connected Vehicles Can Double Throughput in Urban Roads. Transportation Research Part C: Emerging Technologies, Vol. 77, 2017, pp. 292–305.

38.

Smith

S. W.

Kim

Guanetti

Firoozi

Wootton

Kurzhanskiy

A. A.

Borrelli

Horowitz

Arcak

Improving Urban Traffic Throughput with Vehicle Platooning: Theory and Experiments. IEEE Access, Vol. 8, 2020, pp. 141208–141223.

39.

Gordon

Turochy

R. E.

Evaluation of the Effects of Driver Assistive Truck Platooning on Freeway Traffic Flow. Transportation Research Board, Washington, D.C., 2016.

40.

Ioannou

Design and Evaluation of Impact of Traffic Light Priority for Trucks on Traffic Flow. METRANS Transportation Center, Los Angeles, CA, 2015.

41.

Mjogolo

Sando

Soloka

Operational Effectiveness of Truck Priority System: A Microscopic Simulation Approach. Presented at 98th Annual Meeting of the Transportation Research Board, Washington, D.C., 2019, pp. 1–21.

42.

Zhao

Ioannou

A Traffic Light Signal Control System with Truck Priority. IFAC-PapersOnLine, Vol. 49, No. 3, 2016, pp. 377–382.

43.

PTV. PTV Vissim 2021 User Manual. PTV AG, Karlsruhe, Germany, 2021, pp. 1–1278.

44.

Ferguson

Lavery

Higgins

Truck Freight Generators & Attractors in the Province of Ontario. McMaster Institute for Transportation and Logistics, Hamilton, ON, 2014.

45.

Sureshan

Branch

T. P.

Impact of Intercity Trucking on Urban Environment Greater Toronto and Hamilton Area Case Study. Proc., 2009 Annual Conference and Exhibition of the Transportation Association of Canada—Transportation in a Climate of Change, Vancouver, BC, 2009.

46.

MTO iCorridor. Historical AADT & AADTT. Ontario Ministry of Transportation, Toronto, 2021.

47.

MTO. Transportation Tomorrow Survey. Ontario Ministry of Transportation, Toronto, 2016.

48.

Pline

J. L.

Traffic Engineering Handbook, 5th ed. Institute of Transportation Engineers (ITE), Washington, DC, 2000.

49.

Harwood

D. W.

Torbic

D. J.

Richard

K. R.

Glauz

W. D.

Elefteriadou

NCHRP Report 505: Review of Truck Characteristics as Factors in Roadway Design. Transportation Research Board, Washington, DC, 2003.

50.

Van Maarseveen

. Impacts of Truck Platooning at Motorway On-Ramps: Analysis of Traffic Performance and Safety Effects of Different Platooning Strategies and Platoon Configurations Using Microscopic Simulation. Technical Report. TuDelft, Netherlands, 2017.

51.

Yang

Kuijpers

Dane

van der Sande

Impacts of Large-Scale Truck Platooning on Dutch Highways. Transportation Research Procedia, Vol. 37, 2019, pp. 425–432.

52.

Shladover

S. E.

Nowakowski

X. Y.

Ferlis

Cooperative Adaptive Cruise Control: Definitions and Operating Concepts. Transportation Research Record: Journal of the Transportation Research Board, 2015. 2489: 145–152.

53.

Nowakowski

O’Connell

Shladover

S. E.

Cody

Cooperative Adaptive Cruise Control: Driver Acceptance of Following Gap Settings Less than One Second. Presented at Human Factors and Ergonomics Society Annual Meeting, SAGE Publications, Los Angeles, CA, 2010, pp. 2033–2037.

54.

Nowakowski

Shladover

S. E.

Cody

O’Connell

Spring

Dickey

Nelson

Cooperative Adaptive Cruise Control: Testing Drivers’ Choices of Following Distances. University of California, Berkeley, 2011.

55.

Lazar

Tiganasu

Caruntu

C. F.

Arterial Intersection Improvement by Using Vehicle Platooning and Coordinated Start. IFAC-PapersOnLine, Vol. 51, No. 9, 2018, pp. 136–141.

56.

Milanés

Shladover

S. E.

Spring

Nowakowski

Kawazoe

Nakamura

Cooperative Adaptive Cruise Control in Real Traffic Situations. IEEE Transactions on Intelligent Transportation Systems, Vol. 15, No. 1, 2013, pp. 296–305.

57.

VDOT. VDOT VISSIM User Guide. Virginia Department of Transportation (VDOT), 2020.

58.

Zeidler

Buck

H. S.

Kautzsch

Vortisch

Weyland

C. M.

Simulation of Autonomous Vehicles Based on Wiedemann’s Car Following Model in PTV Vissim. Presented at 98th Annual Meeting of the Transportation Research Board, Washington, DC, 2019, pp. 13–17.

59.

NHTSA. Vehicle-to-Vehicle Communication. U.S. Department of Transportation, Washington, D.C., 2020.

60.

Sukennik

Micro-Simulation Guide for Automated Vehicles-D2.5. CoXEist, 2019, pp. 1–29.

61.

Dong

Houchin

Liu

Incorporating the Standstill Distance and Time Headway Distributions into Freeway Car-Following Models and an Application to Estimating Freeway Travel Time Reliability. Journal of Intelligent Transportation System, Vol. 25, No. 1, 2021, pp. 21–40.

62.

Deng

Heavy-Duty Vehicle Platooning—Modeling and Analysis. Doctoral dissertation. KTH Royal Institute of Technology, Stockholm, 2016.

63.

Deng

Boughout

The Impacts of Heavy-Duty Vehicle Platoon Spacing Policy on Traffic Flow. KTH Royal Institute of Technology, Stockholm, 2016.

64.

Ahmed

H. U.

Huang

A Review of Car-Following Models and Modeling Tools for Human and Autonomous-Ready Driving Behaviors in Micro-Simulation. Smart Cities, Vol. 4, No. 1, 2021, pp. 314–335.

65.

WisDOT. Traffic Analysis and Modeling Section 20: Microscopic Simulation Traffic Analysis. In Traffic Engineering, Operations & Safety Manual, Wisconsin Department of Transportation, Wisconsin, 2019, pp. 1–47.

66.

Beeston

Blewitt

Bulmer

Wilson

Traffic Modelling Guidelines V4. Transport for London, UK, 2021.

67.

Dowling

Skabardonis

Halkias

McHale

Zammit

Guidelines for Calibration of Microsimulation Models: Framework and Applications. Transportation Research Record: Journal of the Transportation Research Board, 2004. 1876: 1–9.

68.

Doustmohammadi

Anderson

Kesaveraddy

Jones

Examining Model Accuracy: How Well Did We Do in the 1990’s Predicting 2015. International Journal of Traffic and Transportation Engineering, Vol. 6, No. 1, 2017, pp. 15–21.

69.

FDOT. Traffic Analysis Handbook: A Reference for Planning and Operations. Florida Department of Transportation, Tallahassee, 2014.

70.

Welch

B. L.

The Generalization of ‘STUDENT’S’ Problem When Several Different Population Variances Are Involved. Biometrika, Vol. 34, No. 1–2, 1947, pp. 28–35.

71.

Steenhof

Woudsma

Sparling

Greenhouse Gas Emissions and the Surface Transport of Freight in Canada. Transportation Research Part D: Transport and Environment, Vol. 11, No. 5, 2006, pp. 369–376.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.17 MB