Effect of Traffic Lights Countdown Timer and Motorcycle Lanes as an Approach to the Red Box for Motorcycles in Bali Island

Abstract

The development of red boxes for motorcycles in Indonesia was initially adopted from the advanced stop line (ASL) for bicycles. The bike box concept was adopted for motorcycles in Indonesia. To date, red boxes have been fully implemented in 21 cities in Indonesia. The purpose of this study is to analyze the effect of traffic light countdown timers and motorcycle lanes as an approach to the red box for motorcycles at signalized intersections. There were four locations studied in Denpasar Bali, i.e., red boxes with countdown timer only (Condition 1), red boxes with motorcycle lane only (Condition 2), red boxes with countdown timer and motorcycle lane (Condition 3), and red boxes without countdown timer and without motorcycle lane (Condition 4). The analysis results based on motorcycle volume data indicate that a countdown timer has a significant effect in increasing motorcycle acceleration when the green light starts, reducing the possibility of motorized vehicles other than motorcycles stopping in the red box area and reducing stop line violations while waiting during a red light. Meanwhile, the presence of a motorcycle lane as an approach lane to enter the red box area has a significant influence on increasing the occupancy of the red box by motorcycles. In addition, the correlation test shows that the countdown timer has a strong correlation with the occupancy of the red box to capacity and to stop line violation. Meanwhile, the level of traffic flow is strongly correlated with the countdown timer and motorcycle lane.

1. Introduction

Motorcycles are the most popular mode of transportation and they are widely used by most Indonesians. Indonesia is one of the countries that has the highest motorcycle ownership in the world. Data show that in 2020 the number of motorcycles reached 115,188,762 or approximately 84.5% of the proportion of motorcycles to the total number of motorized vehicles on the road. The growth of motorcycle ownership in 2019 and 2020 was 5.72% and 1.98%, respectively [1,2]. The total population of Indonesia was 270,203,900 at the end of 2020, so motorcycle ownership was 42.63% of the total population of Indonesia. There was a decline in the growth of motorcycle ownership in 2020 due to the COVID-19 pandemic which had an impact on reducing people’s purchasing power.

Indonesian people generally choose motorcycle as their favorite mode of transportation because it has several advantages, such as relatively cheaper price and small dimensions, so that it is able to maneuver on narrow roads, does not require a large parking space, efficient fuel consumption, low vehicle taxes, and low maintenance costs [3]. Motorcycle in Indonesia is a mode of transportation that is used daily, so it has high mobility. Motorcycles are also very popular in other Southeast Asian countries, such as Thailand, Malaysia, Philippines, and Vietnam because motorcycles are proven to be reliable and have financial benefits (i.e., reduced travel costs). The high number of motorcycles is also caused by road conditions, such as many narrow roads, lack of collector roads, and poor road network connectivity, so people’s mobility depends on a flexible and low-cost mode of transportation, i.e., motorcycles [4,5]. In addition, the high number of motorcycles is also due to the ride-hailing service that provides motorcycles with the advantages of cheap transportation costs and fast and reliable travel times [6].

The movement of motorcycles in the signalized intersection during the red-light phase has several characteristics, such as motorcycles that tend to stop in front of other motorized vehicles and the stop line [7]. Motorcycle maneuvers when queuing at a red light have been modeled by considering the dynamic motorcycle lanes, the threshold distance for the motorcycle to start maneuvering, the lane preferred by the motorcycle driver, and the gap acceptance for maneuvering [8]. Furthermore, motorcycle users have a fast response to accelerate their motorcycle when the green light phase begins. The acceleration rate of a motorcycle is higher than a car from the stationery position [9]. Thus, there is a tendency for the irregular movement of motorcycles and potential traffic conflicts with other motorized vehicles.

The development of red boxes for motorcycles in Indonesia was initially adopted from the advanced stop line (ASL) concept for bicycles which has been implemented in several cities in the United States and Canada such as San Francisco, Victoria, Boston, Madison, Portland, Chicago, New York, Austin, and Toronto [10]. The bike boxes, also known as advanced stop lines or advanced stop boxes, were installed to increase the visibility of cyclists and reduce conflicts between motor vehicles and cyclists [11]. The bike box is a promising tool to help bicyclists and motorists avoid conflicts in certain kinds of intersection movements [12]. The green box markings led to significant improvements in bicyclist behavior, but at a considerably higher material cost [13]. The offset stop line was effective in reducing right turn on red conflict with cross street traffic and also resulted in more right turn on red vehicles making a complete stop behind the stop line [14]. The existence of ASLs for bicycles shows an increase in cyclists’ comfort and makes car drivers more aware of the presence of cyclists at signaled intersections [15].

The bike box concept was adopted for motorcycles in Indonesia. The first red box for motorcycles was conducted in June 2007 at the Buah Batu-Soekarno Hatta signalized intersection in Bandung City. The results of red box implementation showed that the level of traffic conflict had decreased and traffic flow had increased at the signalized intersection [16,17]. On the contrary, other studies stated that the red box is less effective, or there is a lot of free space in the back stop line that should be occupied only by motorcycles and at the same time cannot be entered by vehicles other than motorcycles [18]. Furthermore, the application of red boxes is growing very rapidly in Indonesia. To date, red boxes have been fully implemented in 21 cities in Indonesia, such as in Bandung, Bekasi, Tangerang, Bogor, Denpasar, Palembang, Medan, Semarang, Purwokerto, Jepara, Kudus, Cirebon, Bandar Lampung, Kediri, Jember, Jambi, Banda Aceh, Padang, Pemalang, Madiun, and Samarinda. Figure 1 shows the locations of the 21 cities that have implemented red boxes for motorcycles in Indonesia.

ASLs for motorcycles were also developed and applied in other countries. In Malaysia, ASLs for motorcycles are implemented in Kuala Lumpur, namely Zon Motorsikal. The results of the study show that the implementation of ASLs in Kuala Lumpur shows that signalized intersections equipped with ASLs have fewer traffic conflicts compared with signalized intersections that do not have ASLs [19]. ASLs for motorcycles are also applied in Chennai City, India, which shows that applying ASLs for motorcycles can reduce delays at signalized intersections with certain vehicle compositions [20]. Apart from Malaysia and India, ASLs for motorcycles are also widely applied in Taipei City in Taiwan. Taiwan has 14 million motorcycles so it is necessary to give priority to motorcycles and separate them from other motorized vehicles [21]. Furthermore, another previous study stated that there is a close relationship between saturation flow and occupancy rate of ASLs for motorcycles. A low occupancy rate causes a lower saturation rate than without ASLS [18].

Mixed traffic flow at a signalized intersection dominated by motorcycles is very common in most Asian countries such as Indonesia, Malaysia, Thailand, Vietnam, and Taiwan. In Vietnam, the highest proportion of motorcycles in Hanoi and Ho Chi Minh City is 90% [22]. Mixed traffic systems are dynamically complex since there are many parameters and variables that influence the interactions between the different kinds of vehicles [23]. The high number of motorcycle users causes irregular movements of motorcycles while waiting for a red light at a signalized intersection. A high number of motorcycles leads to a large number of traffic violations and a high traffic accident rate [24]. Motorcycles often stop in front of the stop line so that motorcycles block the movement of pedestrians who will cross the road at the zebra cross. The behavior of motorcyclists is categorized as aggressive, normal, and dangerous [25]. Aggressive behavior on motorbikes is often done by young adults [26,27]. Motorcycles will flock to the front of signalized intersections; as a result, many motorcycles will depart together within a very short time once the traffic signal turns green. The mixture of motorcycle traffic will probably enhance the capacity of a street. Motorcycles also often stop in the direct left-turning lane, thus blocking the movement of other motorized vehicles that will turn left [9].

There are significant differences in motorcyclist behavior and accidents between rural and urban areas in Indonesia. Motorcyclists in the urban area are more committed to violations than in the rural area [28]. Improving knowledge is an important factor in decreasing risky motorcycle driving behavior [29]. The behavior of motorcyclists causes a decrease in the performance of signalized intersections. One of the main causes of crash accidents is risky human behavior. This raises the need for proper safety education and training of motorcyclists with traffic law enforcement [30]. At the traffic signalized intersection, they usually stop their motorcycles in front of the stop line and drive through red traffic lights if there are not any vehicles around the junction [31].

To overcome the decline in traffic performance at signalized intersections, it is necessary to carry out a traffic engineering approach, so that performance at signalized intersections becomes more effective by applying red boxes for motorcycles. Signaled intersections are identified as crucial locations for traffic conflicts between motorcycles and other motorized vehicles [32]. The red box is an alternative to overcome the problem of irregular motorcycles at signalized intersections while waiting at a red light. The red box is a designated area at the head of a traffic lane at a signalized intersection that provides motorcyclists with a safe and visible way to get ahead of queuing traffic during the red signal phase. The proportion of idle time significantly influenced fuel consumption and emissions of motorcycles driving on a congested signalized intersection, even with aggressive driving behavior, hard acceleration, and deceleration [33].

By separating motorcycles from four-wheeled vehicles, it is expected that the discharged traffic flow when the green light turns on at signalized intersections will increase. The problem that often arises from the application of the red box for motorcycles is the closed access to the red box because it is blocked by four-wheeled vehicles that stop behind the red box area, blocking the motorcycles from entering the red box area. To overcome this problem, an approach lane for motorcycles was made on the left side to facilitate motorcycles entering the red box area.

In addition, the installation of a countdown timer on a traffic light is expected to reduce the possibility of non-motorcycle vehicles stopping in the red box area when the traffic light turns red. This is because a non-motorcycle driver can find out the remaining green time and the driver can immediately decide whether to stop behind the red box area or speed up the vehicle to pass the red box area. Thus, the existence of a motorcycle lane as an approach to the red box and a countdown timer is considered important to be provided at a signalized intersection where there is a red box for motorcycles.

The purpose of this study is to analyze the effect of traffic light countdown timers and motorcycle lanes as an approach to the red box for motorcycles at signalized intersections. The hypothesis of this study is that the installation of a traffic light countdown timer and the availability of a motorcycle lane as an approach to the red box for motorcycles will have a positive impact on increasing traffic flow, increasing red box occupancy by motorcycles, and reducing stop line violations.

This paper is divided into five sections, i.e., (1) Introduction, (2) Red Box for Motorcycles, (3) Methods, (4) Findings, and (5) Conclusions. The first section contains the background on the application of ASLs for motorcycles which was originally adopted from ASLs for bicycles and subsequently ASLs were developed for motorcycles in Indonesia and other countries. The second section contains a literature review on ASLs for motorcycles in Indonesia and other countries, countdown timers, and motorcycle lanes. The third section contains data collection methods and data analysis methods. The fourth section contains an analysis of traffic flow, occupancy of the red box to capacity, occupancy of the red box by motorcycles and other motorized vehicles, stop line violation, and normality test and analysis of correlation methods. Furthermore, the last section or the fifth section contains the conclusions of the study.

2. Red Box for Motorcycles

Red boxes are a way to manage traffic by separating the queue area between motorcycles and other motorized vehicles while waiting at the red light at an intersection. The existence of red boxes is also meant to provide comfort to the motorcycle driver. Ride comfort is an important aspect for drivers, one of which is due to road profiles [34,35]. Red boxes are placed in front of the stop line for four-wheeled motor vehicles at signalized intersections. Red boxes have two stop lines: one stop line at the front for motorcycles and one stop line at the back for four-wheeled motor vehicles. The two stop line markers are installed sequentially and they are separated by a certain distance approximately 8 m to 12 m.

To implement red boxes, there are geometric requirements and traffic requirements that must be met. First, the geometric requirements for a signalized intersection approach lane are at least a minimum of two lanes, not a direct left turn lane, and the width requirement of each lane is 3.5 m, so that there is a gap for motorcycles to enter the red box area. Second, the traffic condition meets the requirements if there is an irregular queue of motorcycles of at least 20 m long with a minimum number of 30 motorcycles in two lanes while waiting at a red light. Meanwhile, if there are three lanes, the minimum number of motorcycles queueing is 45 motorcycles. Traffic condition requirements are applied in order to encourage the high occupancy of red boxes. Therefore, the implementation of red boxes will be effective [36].

Red boxes equipped with motorcycle lanes as an approach to enter the red box area are located in Jalan Sudirman (South Side) and Jalan Kusumaatmadja in Denpasar, Bali. The motorcycle lanes are 2 m wide. The motorcycle lane is on the left side of the traffic lane with a longitudinal double marking separator, and it is integrated with the red box area. So, motorcycles enter the red box area through the motorcycle lane on the left side. The standard design for motorcycle lanes needs to consider the characteristics of motorcyclists so that the presence of motorcycle lanes can improve road performance [37]. The distance to enter the motorcycle lane needs to be calculated to reduce traffic conflict between motorcycles and other motorized vehicles [38]. The design of the red box for motorcycles that is integrated with the motorcycle lane is shown in Figure 2.

Next, there are also red boxes at signalized intersections with a traffic light countdown timer, which are located on Dewi Sartika road and Kusumaatmadja road. The countdown timer serves to inform motorized vehicle drivers about the remaining time available for each phase so that they can immediately make decisions about when to increase vehicle acceleration, slow down the vehicle, and when to stop. The display on the countdown timer is in seconds. When the green light is on, the driver of the vehicle can know the crucial time before the light turns red. In addition, at a red light, vehicle drivers can get ready before the green light turns on. The countdown timer can also reduce a driver’s boredom while waiting at a red light so that the waiting time does not feel so long. The red box for motorcycles at a signalized intersection that has a traffic light with a countdown timer is shown in Figure 3.

The location of a red box with a countdown timer installed at a traffic light has a tendency to reduce delays or reduce start-up lost time, which is the time lost between the start of the green light and the first vehicle crossing the intersection. The reduced start-up lost time will increase the number of vehicles passing during the green light phase. The effect of installing the countdown timer is at the end of the red light when entering the last three seconds; it should be used to prepare before moving. However, in fact, there are still many drivers in the front row of a traffic light that make an early start even though the light is still red. In addition, at the end of the green light phase when the remaining time is less than three seconds, the vehicle driver should be careful and slow down the vehicle speed. However, the fact is that there are some drivers who actually accelerate the speed of the vehicle, even when the countdown timer shows 0 seconds, and when entering the beginning of the red-light phase, the driver continues to drive through the intersection.

It is important to implement traffic control and synchronize traffic lights that can reduce delays and queue lengths, especially at peak hours [39]. The presence of a countdown timer at a signalized intersection will increase the acceleration and speed of motorized vehicles when the light turns green [40,41,42,43]. The increase in the speed of motorized vehicles occurs during the amber light phase due to the countdown timer, thereby increasing potential traffic accidents [44,45]. Motorcyclists tend to cross the intersection rather than stop during the amber light phase [46].

In addition, the presence of a countdown timer at signalized intersections also reduces red line violation (RLV); [47,48,49,50]. However, the countdown timer will lose its effectiveness in reducing the RLV if the traffic volume is too high [51]. A countdown timer is a type of variable-message sign (VMS). The installation of a VMS on the road can affect driver behavior such as harsh braking, lane changing, and travel speed [52].

The installation of a countdown timer at the red box intersection in Denpasar is intended to facilitate traffic flow. Motorcycles are expected to leave the red box area immediately when the green light is on. In addition, four-wheeled vehicles can predict when the remaining green time available is not sufficient to pass the red box area in the front, thereby reducing the possibility of four-wheeled vehicles suddenly stopping in the red box area for motorcycles. If the red box area for motorcycles is occupied by four-wheeled vehicles, it can reduce the capacity of the red box area for motorcycles. In this study, the presence of a countdown timer was not analyzed using a road safety approach. However, the presence of a countdown timer was analyzed with the traffic flow that passed through the intersection during the green light phase and the red box occupancy for motorcycles.

3. Methods

3.1. Data Collection Methods

The red boxes chosen as locations in this study have different conditions, both in terms of dimensions and facilities. There were four locations studied, i.e., (1) red boxes with countdown timer only, (2) red boxes with motorcycle lane only, (3) red boxes with countdown timer and motorcycle lane, and (4) red boxes without countdown timer and without motorcycle lane. The four locations selected as study locations are shown in

Table 1. Four conditions of red boxes for motorcycles at signalized intersection.

Number	Condition	Intersection	Approach Road	Red Box Dimension			Capacity of Red Box (Number of Motorcycles)
				Width (m)	Length (m)	Area (m2)
1	Red box with countdown timer only	Dewi Sartika-Tengku Umar-Diponegoro	Dewi Sartika	8.08	8	64.64	32
2	Red box with motorcycle lane only	Sudirman-Puputan	Sudirman Road (South)	9.12	8	72.96	36
3	Red box with countdown timer and motorcycle lane	Puputan-Kusumaatmadja	Kusumaatmadja	9.86	8	78.88	39
4	Red box without countdown timer and without motorcycle lane	Sudirman-Puputan	Sudirman Road (North)	9.14	8	73.12	36

. Meanwhile, four conditions of red boxes for motorcycles at signalized intersections are shown in Figure 4.

The data collection method was conducted in three time-phases, i.e., the morning session, afternoon session, and evening session; in each phase of the data collection as many as 10 red-light phases and 10 green-light phases were taken. The total data collected were 30 red-light phases and 30 green-light phases for each red box for motorcycle conditions at signalized intersections. There were four signalized intersection conditions for which data were taken, so the total data collected were 120 red-light phases and 120 green-light phases. The data collection was conducted at peak hours in the morning session, afternoon session, and evening session. The normality test was conducted on the data group or variable, to find out whether the data were normally distributed or not. Meanwhile, the data collected during the green light phase are the volume of motorcycles crossing the intersection. The calculation of the motorized vehicle volume was conducted by surveyors with traffic volume counters.

Data collection was conducted at every signalized intersection condition, i.e., (1) red boxes with countdown timer only (Condition 1), (2) red boxes with motorcycle lane only (Condition 2), (3) red boxes with countdown timer and motorcycle lane (Condition 3), and (4) red boxes without countdown timer and without motorcycle lane (Condition 4). Data collection was conducted in November 2019.

The data collection used a video camera device to record the movement of motorized vehicles at signalized intersections. The video camera was installed on a 6-m-high pole behind the red box area. The distance between the red box area and the camera pole was 20 m, so that the movement of motorized vehicles, especially motorcycles in the red box area, can be monitored properly. Furthermore, the results of the video footage were calculated manually by two personnel.

The volume of motorized vehicles that occupy the red box area during the red-light phase was calculated manually by the traffic counter based on video footage. The application of the red box will be effective if the red box area is only occupied by motorcycles so that no four-wheeled vehicles stop in the red box area. Often the red box for motorcycles is occupied by four-wheeled vehicles, such as passenger cars, so the capacity of the red box for motorcycles is reduced. This makes the red box less effective because motorcycles mix with four-wheeled vehicles when waiting at a red light, so it has the potential to cause traffic conflicts during the green light phase. The low occupancy may be attributed to the misuse of ASL for motorcycles by other vehicles [15,19,53]. To illustrate the condition of a red box during a red light, see Figure 5.

Next, in data collection, the volume of motorcycles that violate the stop lane during the red-light phase was calculated manually by a traffic counter based on video footage. Stop line violation is when motorcycles wait at the red light and stop in front of the stop line. One indicator that shows that the red box is effectively implemented is when no motorcycle stops in front of the stop line. The red box can change the behavior of motorcyclists who initially tend to violate the stop line to not violate the stop line because there is a waiting area for motorcycles in the red box. To illustrate the stop line violation by motorcycles, see Figure 6.

3.2. Traffic Flow Analysis Methods

To analyze whether the traffic flow that occurs at signalized intersections is influenced by the presence of an additional countdown timer and motorcycle lane, a traffic flow analysis was conducted. Data of motorized vehicle flows were collected every 5 s during the green light phase. The results of the data collection were then converted to passenger car equivalent (PCE) per hour. The conversion used a value of 1.3 for large vehicles (trucks and buses), a value of 1.0 for passenger cars, and a value of 0.5 for motorcycles. The PCE value was taken based on the Indonesia Road Capacity Manual [54].

The traffic flow increases from the 0–15th second of green time. After 15 s, the traffic flow tends to be stable until the green time phase ends. At the beginning of the green light phase, a loss of time will occur, where the red-light phase ends and is followed by the beginning of the green light phase. Lost time for motorcycles tends to be more impactful than for passenger cars because motorcycles have a faster initial acceleration than passenger cars [55,56,57]. The placement of motorcycles in front of other vehicles will support the acceleration of motorcycles when the green light starts, thereby reducing lost time.

3.3. Normality Test and Analysis of Correlation Methods

The Kolmogorov–Smirnov normality test is conducted by comparing the distribution of the data to be tested for normality with the standard normal distribution. Standard normal distribution is data that have been transformed into Z-Score form and are assumed to be normal. So, the Kolmogorov–Smirnov test is a test of the difference between the data being tested for normality and standard normal data. Next, if the significance is below 0.05, then there is a significant difference, and if the significance is above 0.05 then there is no significant difference. In the application of the Kolmogorov–Smirnov normality test, if the significance is below 0.05, then the data to be tested have a significant difference from standard normal data, so the data are not normal. Furthermore, if the significance is above 0.05, then there is no significant difference between the data to be tested and the standard normal data. Therefore, the tested data are normal, not different from standard normal.

Simple correlation analysis (bivariate correlation) is used to determine the closeness and direction of the relationship between two variables that occurred. The simple correlation coefficient shows how big the relationship is between two variables. In Statistical Product and Service Solution 25 (SPSS), there are three simple correlation methods (bivariate correlation), namely the Pearson correlation, Kendall’s tau-b, and the Spearman correlation. The Pearson correlation is used for interval or ratio scale data, whereas Kendall’s tau-b and Spearman correlation are more suitable for ordinal scale data.

Simple correlation analysis with the Pearson method is often called the product-moment Pearson. The correlation value (r) ranges from 1 to −1; a value closer to 1 or −1 means the relationship between the two variables is getting stronger, whereas on the other hand, a value close to 0 means the relationship between the two variables is getting weaker. A positive value indicates a direct relationship (X goes up then Y goes up) and a negative value indicates an inverse relationship (X goes up then Y goes down) [58].

4. Findings

4.1. Traffic Flow

Traffic flow data collection was conducted in 30 green light sessions. The average traffic flow based on the four conditions of red box for motorcycle at the signalized intersection is shown in

Table 2. Average traffic flow value in 30 green-light phases.

Period of Time (s)	Red Box with Countdown Timer Only (Condition 1)	Red Box with Motorcycle Lane Only (Condition 2)	Red Box with Countdown Timer and Motorcycle Lane (Condition 3)	Red Box without Countdown Timer and without Motorcycle Lane (Condition 4)
0	0	0	0	0
5	13.18	8.26	14.53	10.09
10	19.02	10.53	19.76	13.84
15	11.19	17.68	11.11	19.34
20	11.04	10.71	8.59	10.25
25	7.27	8.81	5.62	10.34
30	5.42	5.20	4.97	10.09
35	4.13	0.00	0.00	9.42
40	0.00	0.00	0.00	0.00

.

Figure 7 shows the traffic flow when the green light is on at the location of a signalized intersection that has a red box for a motorcycle facility with four conditions. Traffic flow during green time is divided into several periods every 5 s. The traffic flow time periods are 0–5, 5–10, 10–15, 15–20, 20–25, 25–30, and 30–35. The green time in each red box condition is different. The green time in Condition 2 and Condition 3 has a green time of 35 s. In Condition 1 and Condition 4 it has a green time of 40 s.

In Condition 1, which is a red box with a countdown timer, and Condition 3, which is a red box with a countdown timer and motorcycle lane, there is a traffic flow acceleration until the first 10 s after the green time starts. In those conditions, motorcycle volume discharge occurs faster during the green light phase. Meanwhile, in Condition 2, which is a red box with a motorcycle lane, and Condition 4, which is a red box without a countdown timer and motorcycle lane, the traffic flow acceleration does not occur quickly. The peak of motorcycle flow acceleration occurs 15 s after the green phase starts so the motorcycle volume discharge in Condition 2 and Condition 4 does not occur as fast as in Condition 1 and Condition 3.

This shows that the countdown timer has a significant effect on accelerating motorcycles when the light starts to turn green so that the traffic flow increases rapidly within 10 s. After reaching the peak in the 10 s period, then in the 10th until the 40th second the traffic flow tends to decrease. Thus, the installation of a countdown timer on traffic lights can significantly reduce delay or start-up lost time due to traffic flow acceleration in the first 10 s, so that motorcycles can accelerate faster to pass through intersections when the light turns green. Meanwhile, the presence of a motorcycle lane as an approach path to enter the red box does not have a significant impact on increasing traffic flow acceleration during green time.

4.2. Occupancy Red Box for Motorcycles to the Capacity

Maximum occupancy was calculated with a formula where the red box area (m²) is divided by the required area for one motorcycle while waiting at the red box (m²). The required area for one motorcycle was 2 m² (2 m length × 1 m width). Occupancy rate to the capacity is the percentage of the average number of motorcycles that occupy the red box to the maximum red box occupancy. Then occupancy rate by motorcycle only is the percentage of the number of red-light phases occupied by motorcycles only to the total number of red lights.

Based on the data collected on the volume of motorcycles occupying the red box area at the signalized intersection, it was found that the peak volume of motorcycles occupying the red box occurred in the afternoon session, i.e., in the 21st to 30th green time phase. In the afternoon session, the average volume of motorcycles occupying the red box was 28.35 in the four conditions. Condition 3 had the highest volume of motorcycles occupying the red box, which was 36.40. Furthermore, in the afternoon session, the volume of motorcycles occupying the red box reached its lowest number with an average of 15.57 motorcycles. In this morning’s session, Condition 1 had the lowest average volume of motorcycles, which was 12.12 motorcycles. Furthermore, in the morning session, the volume of motorcycles occupying the red box was also quite large, however, it was not as large as the volume of motorcycles in the afternoon session. In the morning session, the average volume of motorcycles occupying the red box was 22.11 motorcycles in all red box conditions. Condition 3 had the highest average volume of motorcycles compared with other conditions, which was 29.6 motorcycles. The volume of motorcycles in three sessions, namely morning session, afternoon session, and evening session during the 30 data collection phases is shown in Figure 8.

Table 3. Red box occupancy to the capacity.

Parameters	Red Box with Countdown Timer Only (Condition 1)	Red Box with Motorcycle Lane Only (Condition 2)	Red Box with Countdown Timer and Motorcycle Lane (Condition 3)	Red Box without Countdown Timer and without Motorcycle Lane (Condition 4)
Red box dimension (m2)	64.64	72.96	78.88	73.12
Red box capacity (motorcycles)	32	36	39	36
Volume of motorcycles/30 red light phase	562	731	871	478
Average red box occupancy (motorcycles)	18.72	24.36	29.04	15.92
Percentage red box occupancy (%)	57.92	67.67	73.63	44.22

shows the highest red box occupancy, namely in Condition 3, red box with countdown timer and motorcycle lane, which is 72.63%. Furthermore, the second highest is in Condition 2, namely red box with motorcycle lane only which has an occupancy rate of 67.67%. However, in Conditions 1 and 4, both of which do not have a motorcycle lane as the approach lane to enter the red box, the occupancy is below 60%. The occupancy of the red box by motorcycles in Condition 1 and Condition 4 is 57.92% and 44.22%, respectively. The results of this analysis indicate that the presence of a motorcycle lane facility as an approach lane to enter the red box area has a significant influence on increasing the occupancy of the red box by motorcycles. This is different from Conditions 1 and 4, where there is no motorcycle lane as an approach lane to enter the red box, so the motorcycle enters the red box through the gap between other motorized vehicles, so there are times when there is a red-light phase where some motorcycles cannot enter the red box area because the gap between passenger cars is too narrow so that the red box area occupancy is not optimal.

4.3. Occupancy Red Box by Motorcycle and Other Motorized Vehicles

The number of red-light phases where the red box area was occupied only by motorcycles in four conditions ranged from 22–25 red-light phases. Conditions 1 and 3 had the highest percentage; 83.3% of the red-light phase was only occupied by motorcycles in the red box area. In Conditions 1 and 3, there is a countdown timer that can reduce the potential for vehicles other than motorcycles to stop in the red box area due to the green light turning red. The presence of a countdown timer can provide a time prediction for the driver to find out the remaining time the light changes from a green light to a red light.

In Conditions 2 and 4 the percentage of red-light phase with the red box area occupied only by motorcycles ranged from 73.33% to 76.67%. This number is lower than Conditions 1 and 3 which have a countdown timer. The type of vehicle that stopped most often in the red box area was a passenger car. Most of the passenger cars that stopped in the red box area were in Condition 2, which was 10 passenger cars out of 30 observed red-light phases. Based on the results of the analysis, the installation of a countdown timer can reduce the potential for stopping vehicles other than motorcycles in the red box area. Furthermore, the occupancy of the red box area by motorcycles is shown in

Table 4. Red box occupation by motorcycle and other motorized vehicles.

Parameters	Red Box with Countdown Timer Only (Condition 1)	Red Box with Motorcycle Lane Only (Condition 2)	Red Box with Countdown Timer and Motorcycle Lane (Condition 3)	Red Box without Countdown Timer and without Motorcycle Lane (Condition 4)
Number of red light phase occupied only by motorcycles in the red box area	25	23	25	22
Percentage of red light phase occupied only by motorcycles in the red box (%)	83.33	76.67	83.33	73.33
Total of non-motorcycles occupied the red box	5	10	6	10
1. passenger car	4	10	6	9
2. bus	1	0	0	1
3. truck	0	0	0	0
Average number of non-motorcycles/red light phase	0.17	0.33	0.20	0.33

.

4.4. Stop Line Violation

The most stop line violations occurred in Condition 4 and Condition 2 by 66 motorcycles and 62 motorcycles, respectively, during 30 red-light phases. The average stop line violation at each red-light phase for Condition 4 and Condition 2 was 2.20 motorcycles and 2.07 motorcycles, respectively. In both conditions, the red box at the signalized intersection does not have a countdown timer. Meanwhile, Conditions 1 and 3, which are red boxes with countdown timers, had lower stop line violations; there were 46 motorcycles and 29 motorcycles for 30 red-light phases, respectively. The average stop line violation at each red-light phase for Conditions 1 and 3 was 1.53 motorcycles and 0.97 motorcycles. This shows that the countdown timer is effective in reducing stop line violations at red boxes at signalized intersections. The stop line violation is shown in

Table 5. Stop line violation.

Parameters	Red Box with Countdown Timer Only (Condition 1)	Red Box with Motorcycle Lane Only (Condition 2)	Red Box with Countdown Timer and Motorcycle Lane (Condition 3)	Red Box without Countdown Timer and without Motorcycle Lane (Condition 4)
Number of stop line violation/30 red light phase	46	62	29	66
Average stop line violation per red light phase	1.53	2.07	0.97	2.20

.

The lowest stop line violations occurred in the afternoon session, which was the 11th to the 20th phase with the number of stop line violations totaling 25 motorcycles in the four red box conditions at signaled intersections. The stop line violations increased by 4.76 times in the evening session, which was in the 21st to the 30th phase. The number of motorcycles violating the stop line was 119 motorcycles. Meanwhile, in the morning session, the number of stop line violations increased by 2.28 times compared with the afternoon session, which was 57 motorcycles. The stop line violation at each green light phase is shown in Figure 9.

4.5. Normality Test and Analysis of Correlation Methods

The occupancy value of the red box capacity that affects each condition based on the normality test, the significance value (p) in Condition 1 (C1) is 0.132, Condition 2 (C2) is 0.200, Condition 3 (C3) is 0.200, and Condition 4 (C4) is 0.200. Overall results show more than the standard value of significance, there is 0.05 (p > 0.05). Based on the significance value (p), the results of the normality test show that the data are normally distributed so it can be continued with parametric statistical analysis, which is shown in

Table 6. Normality test on occupancy value of red box capacity.

Tests of Normality
	Kolmogorov–Smirnov a			Shapiro–Wilk
	Statistic	Df	Sig.	Statistic	Df	Sig.
O_C1	0.132	30	0.193	0.948	30	0.153
O_C2	0.095	30	0.200 *	0.974	30	0.647
O_C3	0.123	30	0.200 *	0.941	30	0.098
O_C4	0.114	30	0.200 *	0.981	30	0.864

* This is a lower bound of the true significance. ^a Lilliefors significance correction.

.

Based on the significance value of Sig. (2-tailed) from

Table 7. Correlation test on occupancy value of red box capacity.

Correlations
		O_C1	O_C2	O_C3	O_C4
O_C1	Pearson Correlation	1	0.776 **	0.824 **	0.655 **
	Sig. (2-tailed)		0.000	0.000	0.000
	N	30	30	30	30
O_C2	Pearson Correlation	0.776 **	1	0.760 **	0.559 **
	Sig. (2-tailed)	0.000		0.000	0.001
	N	30	30	30	30
O_C3	Pearson Correlation	0.824 **	0.760 **	1	0.598 **
	Sig. (2-tailed)	0.000	0.000		0.000
	N	30	30	30	30
O_C4	Pearson Correlation	0.655 **	0.559 **	0.598 **	1
	Sig. (2-tailed)	0.000	0.001	0.000
	N	30	30	30	30

** Correlation is significant at the 0.01 level (2-tailed).

, it is known the value of Sig. (2-tailed) in each condition has a value of 0.00–0.01 < 0.05, which means that there is a significant relationship in each red box condition. Furthermore, based on the calculated r value (Pearson correlations), the largest calculated r value is in the relationship between Conditions 1 and 3, which is equal to 0.824 > r table 0.4093. Both of these conditions have a countdown timer, so it can be concluded that the presence of a countdown timer has a very strong correlation with the red box occupancy to capacity.

Based on the Kolmogorov–Smirnov data normality test in

Table 8. Normality test on stop line violation.

Tests of Normality
	Kolmogorov–Smirnov a			Shapiro–Wilk
	Statistic	Df	Sig.	Statistic	Df	Sig.
SLV_C1	0.212	30	0.150	0.889	30	0.065
SLV_C2	0.206	30	0.081	0.901	30	0.069
SLV_C3	0.218	30	0.072	0.847	30	0.060
SLV_C4	0.129	30	0.200 *	0.923	30	0.033

* This is a lower bound of the true significance. ^a Lilliefors significance correction.

, the significance value for each condition of the influencing stop line violation can be determined, namely C1 is 0.150, C2 is 0.081, C3 is 0.072, and C4 is 0.200. All of the red box conditions have a standard value of significance of 0.05 (p > 0.05). Based on the significance value (p), the data are normally distributed, so it can be continued with parametric statistical analysis.

Table 9. Correlation test on stop line violation.

Correlations
		SLV_C1	SLV_C2	SLV_C3	SLV_C4
SLV_C1	Pearson Correlation	1	0.698 **	0.307	0.638 **
	Sig. (2-tailed)		0.000	0.099	0.000
	N	30	30	30	30
SLV_C2	Pearson Correlation	0.698 **	1	0.344	0.593 **
	Sig. (2-tailed)	0.000		0.063	0.001
	N	30	30	30	30
SLV_C3	Pearson Correlation	0.307	0.344	1	0.395 *
	Sig. (2-tailed)	0.099	0.063		0.031
	N	30	30	30	30
SLV_C4	Pearson Correlation	0.638 **	0.593 **	0.395 *	1
	Sig. (2-tailed)	0.000	0.001	0.031
	N	30	30	30	30

** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed).

shows that based on the significance value of Sig. (2-tailed) in stop line violation, the highest Sig. (2-tailed) value is between Condition 1 and Condition 2, which is 0.00 < 0.05, which means there is a significant relationship. Furthermore, there is a relationship between Conditions 1 and 2, and Conditions 2 and 3 with a sig. value of more than 0.05, which means that there is no significant relationship between these two conditions and the level of stop line violations.

Based on the calculated r value (Pearson correlations), the largest calculated r value is in the relationship between Conditions 1 and 2, which is 0.698 > r table 0.4093. The sig (2-tailed) value obtained is 0.000 (<0.05) so the H0 hypothesis is rejected. This shows that there is a strong correlation between the two variables. Thus, it can be concluded that in Conditions 1 and 2 where there is a countdown timer, it indicates that the stop line violation is affected by the presence of a countdown timer. The presence of a countdown timer can provide traffic light cycle information so that four-wheeled vehicles and motorcycles can stop before the stop line.

Based on the Kolmogorov–Smirnov data normality test in

Table 10. Normality test on traffic flow value.

Tests of Normality
	Kolmogorov–Smirnov			Shapiro–Wilk
	Statistic	Df	Sig.	Statistic	Df	Sig.
TF_C1	0.167	30	0.032	0.880	30	0.003
TF_C2	0.191	30	0.007	0.901	30	0.009
TF_C3	0.160	30	0.049	0.919	30	0.026
TF_C4	0.400	30	0.000	0.644	30	0.000

, the significance value (p) for each variable in each condition for the traffic flow value is 0.032, 0.07, 0.200, and 0.200 for C1, C2, C3, and C4, respectively. All of these data have a standard value of significance, which is 0.05 (p > 0.05). Based on the significance value (p), the Kolmogorov–Smirnov results show that the data are normally distributed except in Condition 4 so that it can be continued with parametric statistical analysis.

Table 11. Correlation test on traffic flow.

Correlations
		TF_C1	TF_C2	TF_C3
TF_C1	Pearson Correlation	1	0.605 **	0.977 **
	Sig. (2-tailed)		0.000	0.000
	N	30	30	30
TF_C2	Pearson Correlation	0.605 **	1	0.634 **
	Sig. (2-tailed)	0.000		0.000
	N	30	30	30
TF_C3	Pearson Correlation	0.977 **	0.634 **	1
	Sig. (2-tailed)	0.000	0.000
	N	30	30	30

** Correlation is significant at the 0.01 level (2-tailed).

shows that the highest significant relationship to the traffic flow value is between Conditions 1 and 3 with a calculated r value (Pearson correlations) of 0.977 > r table 0.4093, and a significance value of 0.000. The other variables that have a positive relationship to the traffic flow value are Condition 2 with Condition 3, and Condition 1 with Condition 2, which are 0.634 and 0.605, respectively, which is greater than the r table value of 0.4487 with a strong correlation level. The Condition 4 variable is not normally distributed so it is not included in the correlation modeling. Thus, it can be concluded that the traffic flow rate in the red box has a significant value in Conditions 1–3, which have a countdown timer and motorcycle lane.

5. Conclusions

The countdown timer is an instrument of variable-message sign, which provides information in real-time to drivers and riders regarding the remaining time on the traffic lights in the red-light phase or the green-light phase. The implementation of the countdown timer supports the concept of smart mobility, especially in the application of red boxes for motorcycles. In order to support smart mobility, the electronic device, i.e., the countdown timer is applied which can increase the effectiveness of the implementation of red boxes for motorcycles.

This study has limitations both in the amount of data taken and the number of intersections. The condition of the signalized intersection in this study was a signalized intersection with a moderate traffic volume, which was not a traffic intersection condition with a high volume. Further research is needed on the effectiveness of the implementation of the countdown timer and motorcycle lane on red boxes for motorcycles with high traffic conditions, especially at signalized intersections in big cities where the number of motorcycles is higher.

Previous research on the effectiveness of ASLs for motorcycles in Indonesia was conducted at signalized intersections without a countdown timer and motorcycle lane. The results of previous research in Indonesia show that the application of ASLs for motorcycles is effective in reducing traffic conflict and increasing traffic flow at signalized intersections [16,17]. Another study shows that the implementation of ASLs is not effective because it leaves empty space behind the stop line due to motorcycles not being able to enter the red box area because they are blocked by four-wheeled vehicles [18]. Meanwhile, the results of ASLs research for motorcycles in Malaysia show that intersections with ASLs show a lower rate of traffic conflict compared with intersections without ASLs [19]. The results of research on the application of ASLs for motorcycles in India show that ASLs can reduce delays at signalized intersections with certain vehicle compositions [20].

This present study fills the gap where the effectiveness of the application of red boxes for motorcycles is associated with the presence of a countdown timer and motorcycle lane as the approach lane to red boxes for the motorcycle area. The results of this study show that the countdown timer has a significant effect on accelerating motorcycles when the traffic light starts to turn green so that the traffic flow increases quickly in 10 s. Meanwhile, the presence of a motorcycle lane as an approach path to enter the red box does not have a significant effect on increasing traffic flow acceleration during green time. The motorcycle lane facility as an approach lane to enter the red box area has a significant influence on increasing the occupancy of the red box by motorcycles. The presence of a countdown timer can provide a time prediction for the car driver to find out the remaining time the light changes from a green light to a red light. If the red box area for motorcycles is occupied by four-wheeled vehicles, it can reduce the capacity of the red box area for motorcycles. Analysis shows that the countdown timer is effective in increasing the occupancy of the red box which is only used by motorcycles. In addition, the countdown timer is effective for reducing the stop line violation in red boxes at signalized intersections.