Evaluating the Effect of Roadside Parking on a Dual-Direction Urban Street

Binghong Pan; Juan Liu; Hu Chai; Yang Shao; Ranyang Zhang; Jiarong Li

doi:10.3791/63384

JoVE Journal > Engineering

Ingénierie

Evaluating the Effect of Roadside Parking on a Dual-Direction Urban Street

Published: January 20, 2023

doi:

10.3791/63384

Binghong Pan¹, Juan Liu¹, Hu Chai¹, Yang Shao², Ranyang Zhang¹, Jiarong Li¹

¹Highway Academy,Chang’an University, ²School of Modern Posts (Logistics School) & Institute of Posts,Xi’an University of Posts and Telecommunications

Summary

In this study, the effect of roadside parking on an urban street is analyzed. The entire process consists of traffic data gathering, data processing, operation simulation, simulation calibration, and sensitivity analysis.

Abstract

Roadside parking is a common traffic phenomenon in China. Narrow urban streets, high parking demands, and a shortage of parking lots force the public to engage in random parking along the roadside. A protocol is proposed to determine the impact of a roadside-parked vehicle on passing vehicles. In this investigation, a dual-direction and two-lane urban street in which one vehicle is parked on the roadside is selected for the collection of traffic data. Based on these data, the impact of the roadside-parked vehicles on the trajectory and speed of passing vehicles is determined. In addition, a microsimulation model is applied to determine the impact of roadside parking on the maximum queue length, delay, emissions, and other indicators under different traffic volumes according to the sensitivity analysis. The results show that roadside-parked vehicles affect the trajectory of passing vehicles for approximately 80 m and have a negative effect on speed, with the lowest speed being observed at the location of the roadside-parked vehicle. The sensitivity analysis results suggest that traffic volume increases synchronously with indicator values. The protocol provides a method for determining the effect of roadside parking on travel trajectory and speed. The research contributes to the refined management of future roadside parking.

Introduction

The acceleration of urbanization is accompanied by an obvious increase in motor vehicle ownership and urban traffic flow. In 2021, China's car ownership reached 378 million, representing an increase of 25.1 million compared to that in 2020¹. However, the current situation with insufficient road capacity and limited traffic management technology has led to an increasingly evident discrepancy between urban traffic supply and demand. Therefore, road traffic congestion has gradually intensified. As the most widespread problem in urban transportation, traffic congestion causes many hazards and has attracted wide attention from researchers²^,³^,⁴. In addition to extending travel time, traffic congestion also aggravates environmental pollution, intensifies energy consumption, and increases pollutant emissions⁵^,⁶^,⁷^,⁸. There is a positive correlation between traffic congestion and accident rates⁹^,¹⁰. Apart from the abovementioned effects, increasing traffic congestion undercuts income and employment¹¹, and this effect is closely related to people's daily life thereby making this one of the main problems in cities. With the development of cities, the adverse impact of road congestion on society will continue to increase.

Traffic congestion is a comprehensive reflection of many urban traffic problems, among which parking is the major one. The expansion of the urban population and the increase in motor vehicles have a negative impact on the parking supply and outstanding parking demand. In the parking system, roadside parking is common in urban traffic and is an important means of addressing the imbalance between parking supply and demand. Roadside parking utilizes resources on both sides of the road to provide parking spaces. Roadside parking is convenient, quick, flexible, and space-saving compared to other parking facilities. However, roadside parking occupies road resources, and its adverse effects cannot be ignored. In cities undergoing rapid development in developing countries, the soaring parking demands make roadside parking overloaded, thus reducing traffic safety, air quality, and public space¹². Therefore, the roadside parking issue needs to be addressed.

Roadside parking space can be located in two scenarios: (1) the non-motorized lane (i.e., on wide roads with separate motorized and non-motorized lanes, roadside parking takes up space on the rightmost non-motorized lane); and (2) the motor vehicle and non-motor vehicle mixed lane, which is often a narrow road with a low traffic volume. As motor and non-motor vehicles share road resources, roadside parking frequently leads to chaos in traffic operations in the second scenario. However, most existing studies have focused on the first scenario¹³^,¹⁴^,¹⁵^,¹⁶^,¹⁷^,¹⁸.

When a roadside parking space is present in the non-motorized lane, and if there is no compulsory isolation of the motorized and non-motorized lanes, roadside parking indirectly leads to mixed traffic. A roadside parking space significantly decreases the effective width of the non-motorized lane, thereby increasing the probability of non-motor vehicles passing through the non-motorized lane and occupying the adjacent motorized lane. The behavior is called lane-crossing¹⁶. Many studies have explored the impact of roadside parking in the non-motorized lane on mixed traffic flow. Based on the cellular automata model, Chen et al.¹³ evaluated the impact of roadside parking on heterogeneous traffic operations in urban streets through the study of friction and congestion conflicts between motor and non-motor vehicles¹³. Chen et al. proposed a road resistance model of mixed traffic flow by considering the effect of roadside parking¹⁷. In addition, some studies have examined the impact of roadside parking only on motor vehicles. Guo et al. proposed a method based on risk duration, which was used to quantitatively analyze the driving time of motor vehicles on roadside parking sections¹⁹, and the results showed that roadside parking significantly impacted travel time.

Traffic simulation is a common tool to investigate the impact of roadside parking. Yang et al. used VISSIM software to explore the impact of roadside parking on dynamic traffic (especially on the capacity), developed a vehicle average delay traffic model, and verified the model reliability through simulation²⁰. Gao et al. analyzed the effect of roadside parking on mixed traffic under four types of traffic interference using the same software¹⁸. Guo et al. used a cellular automata model to analyze the influence of roadside parking on vehicle traffic characteristics (lane capacity and vehicle speed) through Monte Carlo simulation under different scenarios²¹. Under the framework of Kerner's three-phase traffic theory, Hu et al. analyzed the impact of temporary roadside parking behavior on traffic flow based on the cellular automata model²². These studies show that roadside parking has a large negative impact on traffic efficiency.

The traffic management department is interested in understanding the effect of roadside-parked vehicles on the traffic flow. The specific length and degree of the effect are important for managing issues with roadside parking, for example, by providing information on how to delimit parking lots, determine non-parking zones, and regulate parking durations. In this study, a protocol was designed to examine the effect of a single roadside-parked vehicle on traffic operation. The procedure can be summarized in the following steps: 1) preparing the equipment, 2) selecting the data collection location, 3) selecting the investigation time, 4) collecting the data, 5) performing the data analysis, 6) building the simulation model, 7) calibrating the simulation model, and 8) performing the sensitivity analysis. If any requirement in these eight steps is not satisfied, the process is incomplete and insufficient to prove effectiveness.

Protocol

1. Preparation of the equipment

Ensure that all the equipment required is available: radars, roadside laser device, laptops, batteries, a camera, a drone, a reflective tripod, the corresponding cables, and device tripods.

2. Selection of the data collection location (Figure 1)

Select the data collection location. Ensure the location selected is on a two-direction and two-lane road.
NOTE: The location choice is key in this research. The two-lane width is easy to observe.
Ensure that the location does not have any intersections.
NOTE: Vehicles coming from a third direction may cause chaos in the observation.
Ensure that there are no barriers on the road except one parked vehicle placed by investigators.
NOTE: Barriers may interrupt vehicle behavior and block radar detection.
Ensure that there is at least a sight distance and clearance of 300 m. This is required for the radar investigation and the investigators' safety.
NOTE: One radar can detect 200 m at the most. The radars are located 100 m upstream and downstream from the parked vehicle in the investigation.
Ensure that the location is a straight-line segment.
NOTE: If the segment is not straight, it is not possible to determine whether the vehicle offset is caused by the roadside parking.

3. Selection of the investigation time

Select the investigation time. At least 3 h in total is needed, with 1 h during the morning peak, 1 h at the noon, and 1 h during the evening peak²³^,²⁴^,²⁵.
Obtain the time of the peak traffic volume hour from traffic research reports, traffic police departments, or traffic business companies²⁶ (Figure 2).
In the absence of traffic reports or analyses as a reference, collect several hours of data during the three periods above, and then select the data with the highest peak traffic volume²⁷^,²⁸.
Use the hour data with the highest traffic volume to conduct data analysis and as an input into the simulation model. Use all 3 h of data for the model verification.
NOTE: The selected road was flanked by restaurants, and the peak hour for catering is the peak hour for the roadside parking demand. The peak hour for traffic volume is the off-time, and the off-time is also the peak time for catering. Therefore, the peak hour for traffic volume and the peak hour for parking demand are almost synchronous.

4. Data collection (Figure 3)

Park the vehicle approximately 20 cm away from the curb in the intended place so that the roadside laser device can be placed.
Place the reflective tripod at the back of the vehicle. Do not place it too far away to ensure that it does not affect the vehicles' behavior.
NOTE: A reflective tripod is necessary to alert and/or avoid collision based on the relevant provisions of the Chinese Road Traffic Safety Law. The tripod is placed a certain distance behind the parked vehicle to alert the vehicles behind that a parked vehicle is in front and, thus, to avoid a collision. The distance between the reflective tripod and the parked vehicle is kept low in order to minimize the effect of the reflective tripod on the behavior of the passing vehicles, so its effect on the study results is negligible.
Set the radar tripod. Set the tripod at a height of no less than 2 m to avoid signal blockage. Lock the radar with the tripod. Adjust the radar vertically, and turn it toward the parked vehicle. Connect the radar data cable with the laptop USB port.
NOTE: One radar is 100 m upstream and one is 100 m downstream of the parked vehicle. Both radars are placed on the same side of the parked vehicle to catch the traffic data.
Open the radar software, and perform the following steps.
1. Click on Communication Check. Select the Serial Port, and click on Connect. Click on Confirm after the software shows Radar Detected.
2. Click on Investigation Set Up. Click on Read RLU Time and Set RLU Time. Click on Erase Data Record, and confirm it to clear the internal memory of the radar. Click on Start Investigation, and close the dialog box.
3. Click on Real-Time View to check the radar status, and the traffic data should get collected as vehicles pass.
Prepare the roadside laser device and the cable. Connect the roadside laser device data cable with the port. Connect the roadside laser device data cable with the laptop USB port.
Place the roadside laser device in the middle of the parked vehicle. Rotate the four adjustment columns on the device to level it.
NOTE: The roadside laser device must be working under the standard position.
Open the roadside laser device software, and perform the following operations.
1. Click on Communication Check. Select the RLU Serial Port Number, and click on Connect. Click on Confirm after the software shows New RLU Connection Detected.
2. Click on View Investigation. When vehicles pass, the traffic flow will be shown in real time.
3. Click on Investigation Set Up. Click on Read RLU Time and Set RLU Time, successively. Set the Start Time and End Time, and click on Set Task. Click on Confirm after the software shows RLU Investigation Set Up Succeeded.
4. Click on Finish. Click on Device Status to view the status of the roadside laser device.
Set the camera about 30 m upstream of the parked vehicle.
NOTE: The traffic data can be collected by radars and the roadside laser device. Traffic operating videos are prepared for data validation.
Set all the equipment on the double-lane double-side road (here, Dian Zi Yi Road). Check if the radars, roadside laser device, and camera are working well every 5 min.
NOTE: Ensure that the time of the laptops and camera are the same as real time. Start two radars, the roadside laser device, and the camera simultaneously at the scheduled time. Two radars facing each other, combined with an intermediate roadside laser device, provide a continuous trajectory of the affected traffic.
End the data collection, and close the real-time check window in the radar software.
1. Click on Investigation Set Up, select End Investigation, and confirm it. Close the dialog box.
2. Select Data Download, browse the computer to save the data, and input a name for the file. Click on Open, and then click on Start Download. Click on Confirm to finish the radar data collection.
Click on Device Status in the roadside laser device software, and then click on Stop Task to end the data collection. Select Data Download, browse, and input a name for the file. Click on Open, and click on Start Download. Click on Confirm to finish the data collection of the roadside laser device.

5. Data analysis

NOTE: Through data collection, 3 h of data are acquired, including the morning peak, middle noon hour, and evening peak. Playback traffic videos are provided by the camera to calibrate the traffic volumes and vehicle types manually. Select the group data with the highest volume (i.e., the morning peak data in this case) as the representative hour for conducting the data analysis.

Use software to collect the trajectories and speed from the radars.
NOTE: The radar is located 100 m away from the parked vehicle, and the road is 10 m wide. So, all the data points beyond that range are radar errors and should be deleted.
Ensure that the roadside laser device provides the offset value, passing speed, number of vehicles, and the types of vehicles at the parked vehicle position.
Draw the whole range of trajectories and speed provided by the two radars and one roadside laser device as the representative data using calculation software (Figures 4-6).

6. Building the simulation model

NOTE: The microscopic simulation model is established by simulation software for traffic simulation. The results of the data collection, including the traffic volume, vehicle speed, and vehicle type composition, are vital parameters in the traffic simulation and form the basis of the model building. Only the representative data group is needed in the simulation.

Road building
1. Open the simulation software. Import the background map of the investigated road segment.
2. Click on Obstacles on the left, right-click, and select Add New Obstacle. Input the length and width of the obstacle, and then click on OK. Drag the cursor to move the obstacle to the roadway.
  NOTE: "Obstacle" refers to the roadside parked vehicle. The length and width of the obstacle are set according to the actual size of the parked vehicle.
3. Click on Links on the left, move the cursor to the start of the link, and right-click. Select Add New Link, input the lane width, and click on OK. Drag the cursor to draw the link on the map.
4. Repeat step 6.1.3 to build four road segments.
5. Hold the right button of the mouse and the Ctrl button on the keyboard to drag the endpoint of one link to the adjacent link to connect the two links.
  NOTE: This part is called the "connector", and it becomes smoother when more points are added.
6. Repeat step 6.1.5 to connect all the links.
Desired speed
1. Select Base Data from the top bar, and then select Distributions | Desired Speed.
2. Click on the green-cross Add button at the bottom to add a new desired speed distribution and name it.
3. Input the average speed and the maximum speed taken from the representative data as the minimum and maximum desired speeds. Delete the default data.
4. Repeat steps 6.2.2-6.2.3 to establish all the desired speed distributions (the direction of east to west, the direction of west to east, and the reduced speed area).
  NOTE: In the following text, the direction of east to west is abbreviated as E-W, and the direction of west to east is abbreviated as W-E.
Vehicle compositions
1. Select Lists from the top bar, and then select Private Transport | Vehicle Compositions.
2. Click on the green-cross Add button to add a new vehicle composition.
3. Click on the Add button to add two vehicle types: heavy goods vehicles (HGVs) and buses.
4. Select the desired speed distribution set in step 6.2 for cars, HGVs, and buses.
5. Repeat steps 6.3.2-6.3.4 to establish two vehicle compositions (E-W and W-E). Input the flow of cars, HGVs, and buses from the representative data.
Vehicle routes
1. Select Vehicle Routes from the left menu bar.
2. Move the cursor to the upstream of one link, right-click, and select Add New Static Vehicle Routing Decision.
3. Drag the blue cursor to draw the vehicle routes on the map from real routes in the data collection.
Reduced speed areas
1. Select Reduced Speed Areas from the left menu bar.
2. Right-click on the area upstream of the parking position, and select Add New Reduced Speed Area.
  NOTE: The length of the area depends on the data analysis results.
3. Right-click at the margin of the screen, select Add, and select the desired speed set in step 6.2 for the reduced speed area as the area speed.
4. Repeat steps 6.5.2-6.5.3 to set all the reduced speed areas.
Priority rules
1. Select Priority Rules from the left menu bar.
2. Right-click on the reduced speed area upstream of the parked vehicle in the W-E direction, and select Add New Priority Rule. Input the minimum gap time and clearance.
3. Repeat step 6.6.2 to set the priority rule downstream of the parked vehicle in the E-W direction.
  NOTE: The setting of priority rules depends on the real traffic operation reflected by the data collection.
Vehicle travel times
1. Select Vehicle Travel Times from the left.
2. Right-click at the beginning of one link, and select Add New Vehicle Travel Time Measurement.
3. Drag the cursor to the end of the link to build a vehicle travel time measurement.
4. Repeat step 6.7.3 for all the vehicle routes.
Vehicle inputs
1. Select Vehicle Inputs from the left. Right-click at the beginning of one link, and select Add New Vehicle Input.
2. Move the mouse to the left bottom, and input the volume for the representative data.
3. Repeat steps 6.8.1-6.8.2 for all the links.
Nodes
1. Select Nodes from the left. Right-click to select Add New Node, and then click on OK.
2. Left-click and move the mouse to adjust a moderate node range.
  NOTE: The node range is related to the simulation results and depends on the road section geometry.
Click on Evaluation at the top of the simulation interface, and select Result Lists. Click on Nodes Results and Vehicle Travel Time Results.
Click on the blue play button at the top to start the simulation. Click on the device button Quick Mode to maximize the simulation speed.
After the simulation, the node results and vehicle travel time results are shown at the bottom of the interface, including the maximum queue length, parking times, delay, number of vehicles, fuel consumption, CO emissions, NO emissions, VOC emissions, and travel time.

7. Simulation model calibration

NOTE: In this study, the traffic observations showed that the morning peak data had the highest volume, but the three data groups were simulated for verification to fully illustrate the simulation model's reliability.

Input the collected data into the simulation model, run the simulation, and obtain the simulation result (Figure 7A).
NOTE: The simulation volume can be generated from the simulation result.
Compare the simulation volume with the collected volume.
NOTE: Calculate the capacity using Equation 1:
(1)
where C denotes the ideal capacity (veh/h), and h_t denotes the average minimum headway (s).
NOTE: The difference between the collected volume and simulation volume is called the mean absolute percent error (MAPE), as shown in Equation 2:
(2)
where n denotes the four different flows in this study, is the capacity simulated in the simulation model (veh/h), and is the capacity of the investigation (veh/h). The calculated MAPE is listed in Table 2.
NOTE: The simulation accuracy is acceptable when the MAPE is small.

8. Sensitivity analysis

NOTE: Figure 7B shows the sensitivity analysis process. The sensitivity analysis process only reflects the performance of the collected data (Table 3). To understand situations with different traffic volumes in real-time scenarios, all possible traffic volume combinations are input into the simulation model to ensure that all situations are covered in the roadside parking analysis (Figure 8 and Table 4).

Ensure that the representative data contains three groups of data (i.e., W-E volume, E-W volume, and other parameters).
Divide the W-E volume into six categories, divide the E-W volume into seven categories, and keep the other parameters stable in the simulation.
NOTE: The W-E traffic volume was 150-400 veh/h, with an increase of 50 veh/h during the peak hour, and the E-W traffic volume was 150-450 veh/h, with an increase of 50 veh/h during the peak hour. The maximum service traffic volume of one lane in the urban street was 1,140 veh/h.
Simulate 42 situations, and verify the effectiveness under all situations.

Representative Results

This paper presents a protocol to determine the effect of roadside parking on passing vehicles on a two-direction and two-lane urban road through traffic data collection and simulation. A road was selected as the study site (Figure 1), and a vehicle was parked at the planned roadside location. Radars, a roadside laser device, and a camera were applied to collect the vehicle trajectory, speed, volume, and type composition to determine the changes in vehicle trajectory and speed under roadside parking (Figures 4–6). A microscopic simulation model was built based on the geometric characteristics of the road and the data collection results (Figure 7). The sensitivity analysis determined the impact of roadside parking on the maximum queue length, delay, emissions, and other indicators of vehicle operation at varying traffic volumes (Figure 8).

Figure 1 shows the data collection location. The test road was a two-direction and two-lane road in Xi'an city, Shaanxi Province, China. The width of the road was 10 m, the speed limit was 60 km/h, and there was no median strip, representing typical conditions of roadside parking. Two-direction traffic could flow easily but slowed down significantly in the presence of a parked vehicle.

Figure 4 shows the trajectories under the influence of roadside parking based on the data measured by the radars and the roadside laser device. The figure shows that the roadside parked vehicle affected the passing vehicles' trajectory for a length of 80 m. The blue represents the west radar data, and the orange indicates the east radar data. The middle black line is a collection of points, which is the position distribution formed by the vertical position of passing vehicles detected by the roadside laser device.

The west radar shows the trajectory changes. When vehicles saw the roadside parked barrier, they offset from the normal position from 40 m upstream of the parked vehicle.

The roadside laser device could record every passing vehicle's lateral position and speed. The lateral position ranged from 2.3 m to 4.9 m (i.e., the lower and upper ends of the middle black line in Figure 4). The average position was 3.3 m. The position here means the right-side position of the vehicles operating in the W-E direction and the left-side position for the vehicles operating in the E-W direction.

For the east radar, a similar trend to the west radar was observed. Vehicles returned to the normal position approximately 40 m after passing the test vehicle.

As seen in Figure 4, the length of the effect of a roadside parked vehicle on the trajectory of passing vehicles was 80 m. Passing vehicles started to deviate from their normal trajectory at 40 m from the center of the parked vehicle and returned to their normal trajectory after 40 m from the center of the parked vehicle (the exact location is marked with two long black lines in Figure 4, and the horizontal positions of the two lines are 60 m and 140 m). At the parked vehicle position (i.e., the position with coordinates [100,0] in Figure 4), the average distance between the passing vehicles and the outside edge of the parked vehicle was 3.3 m. Considering the width of the parked vehicle, the average distance between the passing vehicles and the inside edge of the parked vehicle was 1.3 m. The minimum and maximum distances between the passing vehicles and the inside edge of the parked vehicle were 0.3 m and 2.9 m, respectively, as determined by the original location and operating conditions of the passing vehicles. Vehicles traveling close to the curb did not have a large lateral distance from the parked vehicle when passing it and even passed close to it at a low speed due to the influence of other vehicles traveling in the same direction. When the passing vehicle was not disturbed by other vehicles traveling in the same direction, the travel width was more generous. In other words, the lateral width between the passing vehicle and the roadside parked vehicle was sufficient. Of course, the lateral width between the passing vehicle and the parked vehicle also depends on driving behavior. Compared with an aggressive driver, a stable driver is more likely to pass a parked vehicle with a greater lateral width.

Figure 5 and Figure 6 show that roadside parking undercut the speed of passing vehicles, with the lowest speed of passing vehicles being observed at the parked vehicle position (i.e., the position with a central horizontal coordinate of [100, 0]). Figure 5 shows the speed in the E-W direction. The traffic is moving from right to left in the picture, indicating that vehicle speed decreases gradually within the 180-120 m range. After passing the parked position, the speed was gradually and evenly distributed without an obvious increase.

In the orange section, just before the parked vehicle position, a maximum speed of 54.7 km/h was reached, and this was the speed at which the vehicle passed at a higher speed than the oncoming vehicle. The lowest speed was 0 km/h, and this occurred at the parking position. With higher offset values of the vehicle in the W-E direction, that vehicle occupied more road width, and the vehicle in the E-W direction had to wait, meaning the latter vehicle's speed was 0 km/h.

In the blue area, after passing the parking position, the vehicle speed stayed in the range of 8-35 km/h. It would have been difficult for cars to reach higher upper speed limits because of the road environment. The lower speed limit slightly increased from 8 km/h to 20 km/h because of driving away from the parking position.

Figure 6 shows the speed in the W-E direction, with vehicles moving from left to right in the picture. The speed changes in the W-E direction were similar to those in the E-W direction.

Before the parking position (i.e., within the range 0-100 m in the figure), the upper and lower limits of vehicle speed in the W-E direction gradually narrowed from the 20 m position. In the range of 0-40 m, the upper limit gradually decreased and was lowest at the 80 m position. The upper speed limit of 38.6 km/h (at the 20 m position) dropped to 29 km/h (at the 80 m position). The lower speed limit increased from 9.4 km/h (at the 10 m position) to 10.44 km/h (at the 100 m position).

The speed limit lowered before the parking position. During the observation, if a vehicle in the W-E direction found the parked vehicle on the same side and there were no vehicles in front of it or the opposite vehicle was far away, the vehicle in the W-E direction tended to accelerate and offset first to occupy a good position for passing the parked vehicle first. This phenomenon is the reason for the speed increase right before the parking position.

Passing the parking position, the speed range was 8.2-47.7 km/h. The lower speed limit decreased because some drivers braked when passing the parked vehicle to avoid scratches. Scratches happen when vehicles coming in both directions meet at the parking spot, and in these cases, drivers attempt to avoid scratches by reducing their speed. Compared with the blue area, the speed limit was increased by 9.1 km/h. This is because when no vehicle was coming in the opposite direction, the vehicles in the W-E direction accelerated by the parking position after confirming that they did not scratch the roadside parked vehicle, consistent with drivers' usual driving habits.

In the orange area, the lower speed limit of 7.5 km/h increased significantly after passing the parking position. This indicates that most vehicles can accelerate back to the speed before the parking position after moving 10 m away from the parking position.

Figure 8 indicates the simulation results of nine indicators that reflect the vehicles' operational status at different traffic volumes. Traffic volumes in E-W and W-E directions affected the maximum queue length (Figure 8A), number of vehicles (Figure 8B), delay (Figure 8C), number of stops (Figure 8D), CO emissions (Figure 8E), NO emissions (Figure 8F), VOC emissions (Figure 8G), fuel consumption (Figure 8H), and travel time (Figure 8I) aligning to the roadside parking data. The increase in traffic volume leads to the increase in all indicator values, but the affected degree of different indicator values is various. Additionally, roadside parking does not have an identical effect on vehicles in E-W and W-E directions.

With the increase in the traffic volume, the impact degree of roadside parking on vehicles in the W-E direction for the three indicators of maximum queue length, delay, and the number of stops was significantly higher than that on vehicles in the E-W direction. In terms of the five emission-related indicators, fuel consumption, and travel time, the impact degree on vehicles in the E-W and W-E directions was almost the same, but it was slightly greater for vehicles in the W-E direction. After the traffic volume reached 300-350 veh/h in the W-E and E-W directions, the growth trend of the maximum queue length, delay, and the number of stops was significantly higher, with the negative impact of roadside parking on the traffic operation efficiency of passing traffic flow becoming more serious. Five of the emission-related indicators, fuel consumption, and travel time changed uniformly with increasing traffic volume in both directions.

Figure 1: The data collection location: a two-direction and two-lane road, Dian Zi Yi Road in Xi'an. Coordinates: 108.932882,34.220774. (A) A schematic of the investigation location in Xi'an City. (B) The red line represents the data collection segment. The north road crossing with the red line is a pedestrian street with few people and does not affect this investigation. Please click here to view a larger version of this figure.

Figure 2: The 24 h congestion index. The data in the panel comes from the real-time congestion index in Xi'an on August 24, 2021²⁶. The data indicate that the morning peak occurred from 07:00 to 09:00 and the evening peak occurred from 17:00 to 19:00. The valley, excluding late night, occurred from 11:00 to 12:00. The congestion indices were 2.25 and 2.66 at 08:00 am and 18:00 pm, respectively. Please click here to view a larger version of this figure.

Figure 3: Data collection scheme shown in a picture taken by a drone at a height of 150 m. Tree shade covers all the equipment, so colorful blocks represent the equipment. The roadside parked vehicle is in the middle, and the two radars are placed 100 m upstream and 100 m downstream of the parked vehicle. The west radar and the east radar both face the parked vehicle. Please click here to view a larger version of this figure.

Figure 4: Whole trajectories. The test vehicle is parked at the position of (100,0) in the panel. The blue represents the west radar data, the middle black line represents the roadside laser device data, and the orange represents the east radar data. Please click here to view a larger version of this figure.

Figure 5: East-West speed. Please click here to view a larger version of this figure.

Figure 6: West-East speed. Please click here to view a larger version of this figure.

Figure 7: Flowcharts for calculating the simulation error (MAPE) and performing the sensitivity analysis. (A) Flowchart for the calculation of the MAPE. (B) Flowchart for the sensitivity analysis. Please click here to view a larger version of this figure.

Figure 8: Sensitivity analysis. X-axis = E-W traffic volume, Y-axis = W-E traffic volume, and Z-axis = evaluation index value. (A) Maximum queue length. (B) Number of vehicles. (C) Delay. (D) Number of stops. (E) CO emissions. (F) NO emissions. (G) VOC emissions. (H) Fuel consumption.(I) Travel time. Please click here to view a larger version of this figure.

Item	Morning (07:00−08:00)		Middle Noon (13:00−14:00)		Evening (17:00−18:00)
Direction	W–E	E–W	W–E	E–W	W–E	E–W
Car(veh/h)	306	374	167	148	351	228
Truck(veh/h)	1	3	1	0	4	0
Bus(veh/h)	9	9	4	5	6	4
Aver. speed(km/h)	21.7	24.5	19.4	24.7	18.8	20.5
Max. speed(km/h)	47.7	54.7	55.8	56.2	44.6	45.0
Min. speed(km/h)	0	0	0	0	0	0

Table 1: Vehicle information collected in the investigation. A minimum speed of 0 km/h indicates that some vehicles stop moving.

Item	Morning (07:00−08:00)		Middle Noon (13:00−14:00)		Evening (17:00─18:00)
Direction	W–E	E–W	W–E	E–W	W–E	E–W
Invest capacity(veh/h)	316	386	172	153	361	232
Simulated capacity(veh/h)	306	360	174	150	354	216
Individual MAPE(%)	3.2	6.7	1.2	2.0	1.9	6.9
MAPE(%)	5.0		1.6		4.4

Table 2: The calibration results for the simulation model. The calibration results between the investigated traffic volume and the simulated volume are listed in the table. The MAPE is calculated using Equation 2, and the errors between the simulated capacity and the actual capacity are 5.5%, 1.6%, and 4.4% for the three data groups, which are all small. As the total capacity error is less than 15%, the error of the established model is within the acceptable range, and the simulation accuracy is sufficient²⁹.

Item	Morning		Middle Noon		Evening
	(07:00−08:00)		(13:00−14:00)		(17:00─18:00)
	W–E	E–W	W–E	E–W	W–E	E–W
maximum queue length(m)	31.26	34.93	12.00	7.96	34.88	20.40
number of vehicles	306	360	168	150	348	216
delay(s)	6.47	6.58	3.10	1.74	6.68	4.64
number of stops(times)	0.28	0.52	0.05	0.11	0.24	0.42
CO emissions(grams)	191.790	249.606	89.112	77.820	219.462	135.468
NO emissions(grams)	37.314	48.564	17.340	15.138	42.702	26.358
VOC emissions(grams)	44.448	57.846	20.652	18.036	50.862	31.398
fuel consumption(gallon)	2.742	3.570	1.272	1.116	3.138	1.938
travel time(s)	35.46	29.12	31.92	24.56	35.73	27.25

Table 3: Simulation results with the morning peak data, middle noon data, and evening peak data. As the representative data, the morning peak data group has the highest traffic volume and indicator values. The middle noon traffic data group has the lowest traffic volume and indicator values.

Item	Value
E–W volume(veh/h)	150/200/250/300/350/400/450
W–E volume(veh/h)	150/200/250/300/350/400
Note:The E–W traffic volume is in the range 150–450 veh/h with an increase of 50 veh/h. The W–E traffic volume is in the range 150–400 veh/h with an increase of 50 veh/h.

Table 4: Input parameters for the sensitivity analysis in the simulation.

Discussion

The effect of roadside parking on urban streets cannot be ignored, and random parking needs to be addressed³⁰^,³¹. A protocol to determine the impact of roadside parking on traffic flow in a dual-direction urban street is presented here. The data collection specifies the trajectory and speed changes of passing vehicles caused by roadside parking. The traffic simulation quantifies roadway indices such as maximum queue length, delay, and emissions.

The critical steps in the protocol are the data collection and the microsimulation model building. The data collection location is a straight segment without an intersection, entrance, or exit. To ensure the influence is visible, the road cannot be wider than 10 m. A street with a width of 10 m is right for observation. If narrower, the traffic may be broken down completely, and if wider, the influence may not be detected. A long enough sight distance is also a requirement for the segment. When establishing the simulation model, attention should be given to reduced speed areas and priority rules. The relevant parameters (speed and length) of the reduced speed areas are set based on the representative data to reflect the actual road operation. The drivers’ behavior can be better reflected using priority rules instead of conflict areas. The priority rules are the same as the representative data and are checked using the traffic operation videos captured by the camera.

Regarding the effect of roadside parking on passing vehicles, this protocol provides a specific and realistic description of the investigation results. For example, the trajectories of passing vehicles are affected for 80 m in length, and vehicle speed is also negatively impacted. In addition, under different traffic volumes, the simulation analysis results show the performance of various indices that reflect the traffic operation efficiency. The increase in traffic volume is synchronous with the growth in the indicator values.

The main limitation of this protocol is that it is only effective for one parked vehicle on the roadside. The next research stage will be conducted to determine the effect of multiple randomly parked vehicles on traffic flow operation.

It is recommended that traffic police add monitoring equipment on narrow urban streets to monitor vehicles parked on the roadside thereby mitigating the impact of roadside parking.

The protocol described here for evaluating the effect of roadside parking on a dual-direction urban street can be applied to propose refined roadside parking management measures, such as the allowable parking time, the recommended parking location, and the permissible parking vehicle types.

Divulgations

The authors have nothing to disclose.

Acknowledgements

The authors would like to acknowledge the Scientific Research Program Funded by Shaanxi Provincial Education Department (Program No. 21JK0908).

Materials

battery	Shenzhen Saiqi Innovation Technology Co., Ltd	LPB-568S
cables for radar	BEIJING AOZER TECH & DEVELOPMENT CO.,LTD
cables for roadside laser device	MicroSense
camera	Sony Group Corp	HDR-CS680
camera tripod	Sony Group Corp
drone	SZ DJI Technology Co.,Ltd.	DA2SUE1
laptop	Dell	C2H2L82
radar	BEIJING AOZER TECH & DEVELOPMENT CO.,LTD	CADS-0037
radar tripod	BEIJING AOZER TECH & DEVELOPMENT CO.,LTD
reflective tripod	Beijing Shunan liandun Technology Co., Ltd
roadside laser device	MicroSense

References

He, Y. X. A traffic capacity model of lane occupation. Applied Mechanics and Materials. 599-601, 2083-2087 (2014).
Hua, S. Y., Wang, J. L., Zhu, Y. Cause analysis and countermeasures of Beijing city congestion. Procedia-Social and Behavioral Sciences. 96, 1426-1432 (2013).
Yang, H. X., Li, J. D., Zhang, H., Liu, S. Q. Research on the governance of urban traffic jam based on system dynamics. Systems Engineering-Theory & Practice. 34 (8), 2135-2143 (2014).
Rajé, F., Tight, M., Pope, F. D. Traffic pollution: A search for solutions for a city like Nairobi. Cities. 82, 100-107 (2018).
Abdull, N., Yoneda, M., Shimada, Y. Traffic characteristics and pollutant emission from road transport in urban area. Air Quality, Atmosphere & Health. 13 (6), 731-738 (2020).
Shi, K., Di, B. F., Zhang, K. S., Feng, C. Y., Svirchev, L. Detrended cross-correlation analysis of urban traffic congestion and NO 2 concentrations in Chengdu. Transportation Research Part D: Transport and Environment. 61, 165-173 (2018).
Lu, Q. Y., Chai, J., Wang, S. Y., Zhang, Z. G., Sun, X. C. Potential energy conservation and CO2 emissions reduction related to China’s road transportation. Journal of Cleaner Production. 245, 118892 (2020).
Sánchez González, S., Bedoya-Maya, F., Calatayud, A. Understanding the effect of traffic congestion on accidents using big data. Sustainability. 13 (13), 7500 (2021).
Fuente, J., Rolloque, A. C., Azas, P., Alcantara, M. M. Young road safety advocate program, the "peer to peer" approach in teaching pedestrian safety. Injury Prevention. 22, 67 (2016).
Jin, J., Rafferty, P. Does congestion negatively affect income growth and employment growth? Empirical evidence from US metropolitan regions. Transport Policy. 55, 1-8 (2017).
Ajeng, C., Gim, T. Analyzing on-street parking duration and demand in a metropolitan city of a developing country: A case study of Yogyakarta City, Indonesia. Sustainability. 10 (3), 591 (2018).
Chen, J. X., Li, Z. B., Jiang, H., Zhu, S. L., Wang, W. Simulating the impacts of on-street vehicle parking on traffic operations on urban streets using cellular automation. Physica A: Statistical Mechanics and its Applications. 468, 880-891 (2017).
Ye, X. F., Chen, J. Impact of curbside parking on travel time and space mean speed of nonmotorized vehicles. Transportation Research Record. 2394 (1), 1-9 (2013).
Ye, X., Yan, X. C., Chen, J., Wang, T., Yang, Z. Impact of curbside parking on bicycle lane capacity in Nanjing, China. Transportation Research Record. 2672 (31), 120-129 (2018).
Guo, H. W., Gao, Z. Y., Zhao, X. M., Yang, X. B. Traffic behavior analysis of non-motorized vehicle under influence of curb parking. Journal of Transportation Systems Engineering and Information Technology. 11 (1), 79-84 (2011).
Chen, J., Mei, Z. Y., Wang, W. Road resistance model under mixed traffic flow conditions with curb parking. China Civil Engineering Journal. (09), 103-108 (2007).
Gao, L. P., Sun, Q. X., Liu, M. J., Liang, X., Mao, B. H. Delay models and simulation on mixed traffic system with curb parking. Journal of System Simulation. 22 (003), 804-808 (2010).
Guo, H. W., Gao, Z. Y., Yang, X. B., Zhao, X. M., Wang, W. H. Modeling travel time under the influence of on-street parking. Journal of Transportation Engineering. 138 (2), 229-235 (2012).
Yang, X. G., Long, L., Pu, W. J. Optimal distance between one-side curbside parking location and signalized intersection. Journal of Tongji University (Natural Science). 33 (3), 297-300 (2005).
Guo, H. W., Wang, W. H., Guo, W. W. Micro-simulation study on the effect of on-street parking on vehicular flow. 2012 15th International IEEE Conference on Intelligent Transportation Systems. , 1840-1845 (2012).
Hu, X. J., Hao, X. T., Wang, H., Su, Z. Y., Zhang, F. Research on on-street temporary parking effects based on cellular automaton model under the framework of Kerner’s three-phase traffic theory. Physica A: Statistical Mechanics and its Applications. 545, 123725 (2020).
Shao, Y., et al. Evaluation of two improved schemes at non-aligned intersections affected by a work zone with an entropy method. Sustainability. 12 (14), 5494 (2020).
Shao, Y., et al. Evaluating the sustainable traffic flow operational features of an exclusive spur dike U-turn lane design. PLoS One. 14 (4), 0214759 (2019).
Shao, Y., Han, X. Y., Wu, H., Claudel, C. G. Evaluating signalization and channelization selections at intersections based on an entropy method. Entropy. 21 (8), 808 (2019).
Xi’an realtime traffic congestion delay index. AutoNavi Traffic Big-data Available from: https://trp.autonavi.com/detail.do?city=610100 (2021)
Pan, B. H., et al. Evaluation and analysis model of the length of added displaced left-turn lane based on entropy evaluation method. Journal of Advanced Transportation. 2021, 2688788 (2021).
Pan, B. H., et al. Evaluating operational features of three unconventional intersections under heavy traffic based on CRITIC method. Sustainability. 13 (8), 4098 (2021).
Sun, J. . Guideline for Microscopic Traffic Simulation Analysis. , (2014).
Koohpayma, J., Tahooni, A., Jelokhani, N. M., Jokar, A. J. Spatial analysis of curb-park violations and their relationship with points of interest: A case study of Tehran, Iran. Sustainability. 11 (22), 6336 (2019).
Zoika, S., Tzouras, P. G., Tsigdinos, S., Kepaptsoglou, K. Causal analysis of illegal parking in urban roads: The case of Greece. Case Studies on Transport Policy. 9 (3), 1084-1096 (2021).

Play Video

PDF

DOI

DOWNLOAD MATERIALS LIST

Citer Cet Article

Pan, B., Liu, J., Chai, H., Shao, Y., Zhang, R., Li, J. Evaluating the Effect of Roadside Parking on a Dual-Direction Urban Street. J. Vis. Exp. (191), e63384, doi:10.3791/63384 (2023).