Priority-Setting Methodology of Qualitative and Quantitative Analyses for Pedestrian Road Construction: Case Study of a National Highway in South Korea

Abstract

This study aimed to propose a priority-setting methodology for the construction of pedestrian roads using a set of evaluation methods distinguished from the conventional system using social overhead capital(SOC) evaluation with a focus on economic analysis. Four evaluation indicators were selected in this study: pedestrian fatality risk, surrounding infrastructure, potential pedestrian demand, and manager’s opinion, with a total of 12 attributes across the indicators. Each attribute was designed to allow a quantitative or qualitative evaluation and had a method of estimating and scoring each attribute for set criteria. Regarding the weighting of evaluation indicators, the Analytic Hierarchy Process was applied as the most widely used method in multi-criteria decision making. The highest weightage was placed on pedestrian fatality risk and surrounding infrastructure among the evaluation indicators and traffic accidents of pedestrian fatality risk among the attributes. To determine the practical application of the proposed methodology, a single-site evaluation was conducted, followed by 385 sites in a similar manner for priority setting. The methodology of priority setting for pedestrian roads proposed in this study can be applied to the priority setting for pedestrian roads, smaller-scale SOC projects including bicycle roads.

1. Introduction

Road traffic policies in South Korea have advanced with a focus on vehicles to support growth-centered industrialization, leading to social problems such as severe traffic congestion in urban areas, environmental pollution, and energy depletion. To resolve road traffic problems, the government has implemented green transportation policies since the 2000s, and with increased interest in safety, environment, and quality of life in South Korea and overseas, the interest and relevant research of pedestrian travel for green transportation have increased [1]. Pedestrian travel is an essential element of all transportation means and is the primary means available for older adults, children, and the physically disabled, who constitute the vulnerable transportation group. Notably, in South Korea, where the era of aged and low-birth society is fast approaching across the globe, it is crucial from the perspective of social equity to enhance pedestrian safety for this vulnerable group. Thus, the government has enacted the criteria for designing pedestrian roads and developed the basic plan of nationwide pedestrian road construction to improve pedestrian safety on national highways in each region, and accordingly, pedestrian roads have been constructed [2].

National highways in South Korea consist of 52 lines as of 2020, totaling 14,098 km. These roads function as the routes connecting the major cities, ports, airports, and touristic sites across the nation and serve as the national motorway network. However, since the 1990s, the function of national highways has been biased towards the role of motorways where accessibility is not adequately reflected; hence, the residents of the regions close to national highways are not guaranteed the right to pedestrian safety. Notably, the level of pedestrian safety has been substantially reduced at sites without pedestrian roads in the regions close to national highways passing directly through the villages whose residents require highway access for key facilities such as public offices and public transportation, including buses as shown in Figure 1.

According to the relevant law, pedestrians on national highways must use the lane shoulder at a site without a pedestrian road. However, at sites without ≥1.5 m lane shoulder width, the vehicles running at high-speed pose difficulties to pedestrian travel. Hence, the government has continuously executed projects to construct pedestrian roads for safety on national highways since the 2000s. Additionally, various projects are continuously executed at sites without pedestrian roads, such as increasing the lane shoulder width to enhance pedestrians’ safety. Nonetheless, the number of vehicles to pedestrian accidents on national highways has increased from 1897 cases in 2017 to 1979 in 2021. In contrast, the number of vehicles to pedestrian accidents on all road types has significantly decreased from 46,728 cases in 2017 to 35,269 in 2021. This demonstrates that the level of pedestrian safety on national highways remains low. The mortality, as a representative measure of the risk of traffic accidents, was reduced from 11.6 in 2017 to 7.4 in 2021 for national highways, but the figure was two-fold higher than provincial roads (4.5) or urban roads (2.9) [3]. This implies the lack of infrastructure, such as pedestrian roads, to ensure the safety of pedestrians, while the vehicle running speed on national highways is far higher than that on other roads, including provincial and urban roads.

In South Korea, the total length of existing pedestrian roads was approximately 490 km on national highways in 2019, and even after considering the total length of all national highways at 14,000 km and that specific sites such as vehicle-only roads prevent the construction of pedestrian roads, the current number of pedestrian roads is deficient. Notably, there are many sites on national highways without pedestrian roads, except those that pass through residential villages in rural areas. However, due to the limited budget, building pedestrian roads at all sites on national highways is impossible, and a practical solution would be to construct pedestrian roads on selected sites that demand prioritized construction one at a time. A study should be conducted to investigate selecting the sites that demand prioritized construction of pedestrian roads.

Various methods can be considered in setting the priority of pedestrian road construction among such sites with high demand, including quantitative and qualitative methods. The most straightforward method of quantitative estimation is to quantify the number of traffic accidents. However, the method is not ideal, considering that most traffic accidents have human factors. Thus, instead of a simple indicator like a traffic accident, various attributes related to the pedestrian fatality risk must be considered. Additionally, because pedestrian travel as a mode of transportation is characterized by the connectivity to visiting key facilities or transit across public transportation, such specific attributes should also be incorporated. Moreover, the traffic infrastructure should fundamentally reflect the transportation demand, and to set the priority for the construction of pedestrian roads, the pedestrian travel demand should also be incorporated. Lastly, in contrast to traffic roads, pedestrian roads are constructed within one-three km lengths to maximally reflect the opinions of those involved in candidates’ management. While quantitative attributes are necessary to ensure objectivity across the evaluation indicators, qualitative judgments by the person undertaking the evaluation are also required for such factors because the connectivity to key facilities cannot be incorporated through quantitative attributes.

Hence, this study aimed to propose a methodology involving qualitative and quantitative means of priority setting for the construction of pedestrian roads that reflect the characteristics of national highways. For this, the qualitative or quantitative evaluation indicators and attributes were defined and followed by developing a respective evaluation method for each attribute. Additionally, the method of weighting each attribute and the methodology of priority setting, which can be applied to the sites that demand the construction of pedestrian roads, was explored. The practical application of the proposed methodology was determined by collecting the data from the candidate sites on an actual national highway and applying the evaluation indicators methods and attributes to set the priority on each candidate site. The novel methodology can be applied in setting the priority for pedestrian road construction on national highways and all roads by the government, including provincial roads.

2. Literature Review

Various studies have been conducted in South Korea and overseas regarding increasing pedestrian safety. The existing research on the evaluation of pedestrian facilities mainly focuses on there aspects: the characteristics of pedestrians’ and traffic accident analysis and LOS (Level of Service) analysis. Kwon et al. [4] analyses the effect of pedestrians’ personal and household characteristics and accessibility to public facilities on a walking safety satisfaction level. As a result of the analysis, age and gender variables were important for elderly people, and education was important factor for disabled people. In addition, there are many other studies regarding the characteristics of pedestrians [5,6,7,8,9,10]. Traffic accident analysis is one of the most basic methods when evaluating road safety. Jun et al. [11] predicted elderly and child pedestrian hot spots in the Seoul Metropolitan Ares. They used the traffic accident analysis system data and collected Google Street View images of elderly and child pedestrian crash. Additionally, there are numerous studies have frequently been conducted [12,13,14,15,16,17]. Additionally, for pedestrian road evaluation, numerous studies have investigated the method of estimating the level of service (LOS) [18,19,20,21].

While the evaluation and LOS analysis of existing pedestrian roads and safety facilities is conducted, few studies have explored suitable locations for constructing new pedestrian roads. Anderw R.S [22] made a list of priority bicycle corridors and pedestrians locations, which are reviewed within the collaborative metropolitan planning organizations planning forum. Two evaluation indicators were selected to prioritize the installation of priority of pedestrian road: Suitability and Demand. The suitability includes a traffic volume and speed, number of travel lanes and shoulder and so on. The Demand include a population and employment, transit accessibility and so on. Using the proposed indicators, they proposed the installation priorities of New Jersey as High Priority Corridor, Medium Priority Corridor, Low Priority Corridor.

Jeon, W.H [2] proposed methodology to prioritize sidewalk construction three main steps: (1) analytic Hierarchy Process methods, (2) Subjective evaluation of relevant road agencies for the candidate sidewalks, and (3) Field study conduction. Methodology of this study included survey data and expert consultation. In order to demonstrate the reasonableness of the proposed methodology, a case study was performed for exactly 100 candidate site for sidewalk construction. In addition, the city of Kirkland proposed the factors that entered into the walkway evaluation [23]. Four evaluation indicators were selected to prioritize the installation of priority of pedestrian road: a distance from site to school, Transit, park and commercial area. Each of indicators are weighted; School and parks at 30% and Transit and commercial area at 20% for a total of 100%.

The study of Anderw et al. [22] is meaningful in reflecting field data such as traffic volume and number of lanes that affect pedestrian safety. Additionally, it is valuable in including indicators of transit accessibility such as population and public transportation station. However, it is judged that the indicators for priority selection are relatively simple and the opinions of road managers are not reflected, so the indicators for actual needs are insufficient.

The study of Jeon et al. [2] is shows that the evaluation indicators are relatively well presented as safety, demand generation, and connectivity, but the limitation is that the detailed indicators of demand generation are proposed as pedestrian traffic that is difficult to investigate in the field. The study [23] of Kirkland city suggested a methodology for the priority of installation of pedestrian roads through various evaluation indicators, but it is judged that There is a lack of evidence for the weights applied.

3. Methods

3.1. Procedures of Priority Setting

Pedestrian roads are mainly constructed on the parallel side to traffic roads, except in park or river areas; hence, the evaluation indicators and methods vary according to the type and characteristics of the traffic roads in the vicinity. Even on national highways, the demand for pedestrian road construction to ensure pedestrian safety may increase if close to residential villages. In contrast, at national highway sites with high design speed and very low demand of and accessibility by pedestrians, the need for pedestrian road construction decreases. Additionally, the road types with a greater emphasis on accessibility than mobility, such as city and county roads, wherein the basic pedestrian demand is guaranteed, pedestrian fatality risk and connectivity to surrounding infrastructures is a more critical evaluation indicator. Thus, in planning pedestrian road construction, the type and characteristics of the traffic road are essential factors to be considered.

Once the traffic road type and evaluation indicators are determined, the respective evaluation methods should be developed. Each method should offer an adequate explanation for the given evaluation indicator, which may be quantitative or qualitative. A quantitative evaluation method should allow on-site measurements, and a qualitative one should provide an accurate set of evaluation criteria to prevent scoring bias.

After defining the evaluation indicators and methods, the weighting of each indicator is a necessary step. The weight of an evaluation indicator is expressed by the factors considered in pedestrian road construction, which may vary in each evaluator. Thus, the weighting should be conducted following a method of multi-criteria decision-making. In the survey used in weighting, clear definitions should be given for the road type and evaluation indicators.

After evaluation indicators, methods selection, and weighting is completed, a list of sites requiring pedestrian road construction and respective field investigations should be conducted. Given that pedestrian roads are mostly short, unlike traffic roads, the opinions of relevant managers should be maximally reflected. Hence, the opinions of the management agency of the candidate site should be reflected on the list, while the field investigations should target all sites. Afterward, the result should be quantified through scoring based on the weight of each evaluation indicator, and based on the scores, priority can be set on the candidate sites from the one with the highest demand. The procedures, from selecting the road type to determine the priority regarding the construction of pedestrian roads to analyzing the priority, are outlined in Figure 2.

The methodology to determine the priority for the construction of a pedestrian road varies according to the road type for the selection of evaluation indicators, methods, and weighting. This study aimed to define a methodology of priority setting for the construction of pedestrian roads on national highways among the seven road types (Expressway, National Highway, Special metropolitan city roads, Provincial Road, City Road, County Road, District Road) stated in the Road Act in Korea. While national highways are the type of road with an emphasis on mobility to connect across cities, they are simultaneously characterized by the accessibility by residential villages in rural areas. Thus, the proposed methodology of priority setting for national highways would be applicable to other road types.

3.2. Methodology of Priority Setting for the Construction of Pedestrian Roads on National Highways

3.2.1. Selection of Evaluation Indicators

Objective and clear data and evaluation indicators are needed to select the priority of road infrastructure installation such as pedestrian roads. However, as suggested in a report by the world bank [24], governments often misjudge or waste unnecessarily budgets under limited data and evaluation methods when selecting priorities. In order to make the best judgment under limited conditions, this study attempts to apply the stepping stone approach. There are three stages of the stepping stone approach according to the World Bank report, and in this study, a second-stage approach was applied, as shown in Figure 3. The indicators proposed in this study is an example that can be applied to the installation of pedestrian roads in Korea, and may vary depending on the country to be applied.

In the priority setting for the construction of pedestrian roads on national highways, the evaluation indicators should be selected but evaluation indicators can be modified depending on the road type or traffic conditions. Therefore, in this study, four indicators were selected in consideration of the situation in Korea as follows. First, due to the characteristic high vehicle running speed on national highways, a factor influencing pedestrians’ safety is included. Second, a factor related to the trip generation and assessment reflecting the pedestrian demand is included. Third, to maximize the economic feasibility, the connectivity to the existing pedestrian roads, public transportation, public facilities, and touristic sites must be considered. Lastly, the opinions of those in charge of the management of the candidate site, which can reflect the need for pedestrian road construction, should be reflected. A schematic diagram of the steps in the flow toward pedestrians is given in Figure 4.

Regarding a factor influencing pedestrians’ safety, an attribute to indicate the pedestrian fatality risk is included. Such factors on national highways include the lane shoulder width, the vehicle speed limit, traffic volume, and the number of traffic accidents. In the absence of a pedestrian road on a national highway, the lane shoulder should be used by pedestrians. With adequate lane shoulder width, the priority of pedestrian road construction would be low. In contrast, if the lane shoulder width is below the legal criteria, a priority could be assigned. The vehicle speed limit and traffic volume are the two most direct risk factors for pedestrians using the lane shoulder for travel. With high levels of the vehicle speed limit and traffic volume, pedestrians would inevitably be constantly exposed to vehicles, reducing pedestrians’ safety. Lastly, the number of traffic accidents is the most representative quantitative attribute to indicate traffic safety at a candidate site. Thus, the first evaluation indicator in the priority setting for the construction of pedestrian roads on national highways is pedestrian fatality risk, whose attributes are the mean lane shoulder width, speed limit, average annual daily traffic (AADT), and traffic accidents.

Regarding the factor reflecting pedestrians’ demand, an attribute to indicate pedestrian population and pedestrian-generating facilities is included. First, regarding pedestrian population, a strong influencing factor is the number of residents in a region close to the candidate site of pedestrian road construction. However, as the means of pedestrian travel vary across age groups, the rate of pedestrians in each age group among the residents should be considered. For pedestrian-generating facilities, while it could be estimated based on the demand of visiting a public or commercial facility at the candidate site, such data is non-existent at the concerned facilities. Thus, this study defined pedestrian-generating facilities based on the number of public and commercial facilities in the vicinity of the candidate site. Thus, the second evaluation indicator is the potential pedestrian demand, whose attributes are the pedestrian population and pedestrian-generating facilities.

Connectivity to pedestrian roads, traffic facilities, public facilities, and touristic sites is the indicator that can increase the economic value of the pedestrian road to be constructed. To illustrate, if a pedestrian road could connect two isolated areas, the economic value of the pedestrian road could be enhanced at a low cost. Likewise, if a pedestrian road is built in an area on the way to a public facility, a public transportation facility or a tourist site, it can cause significant economic gains. Thus, the third evaluation indicator is surrounding infrastructure, whose attributes are connectivity to pedestrian roads, connectivity to key facilities, connectivity to traffic facilities, and tourism.

Lastly, as pedestrian roads are a type of public facility that increase citizens’ convenience built at a site of a relatively short distance, an indicator to reflect on-site opinions to determine the need for pedestrian road construction is necessary. The optimum method is to conduct a survey on the residents in the vicinity of the candidate site, but due to the characteristics of the SOC, the survey result is likely to be biased towards necessity across almost the entire group of residents. Thus, in reflecting on-site opinions, it is necessary to collect the opinions from road managers with a high level of understanding of the issues around the candidate site rather than from pedestrian road users. However, because managers’ opinions may still be biased towards necessity, the priority should be assigned upon request for multiple sites at each respective road management agency. Additionally, the need for the construction, such as civil complaints and traffic risks generated at the candidate site, must be examined and quantified. Thus, the last of the evaluation indicators is the manager’s opinion on the priority at the candidate site with construction rationale.

The four indicators and attributes presented in this study were composed of independent indicators and attributes, but some attributes may be related. Therefore, the indicators and attributes presented in this study can be changed or added/deleted in other studies. In addition, the method of classifying grades for each evaluation indicators was to follow national standards or reference such as shoulder width or speed limit. On the other hand, evaluation indicators without national standards or reference were classified using the point of inflection or the same interval of the data used in the case study of this study.

3.2.2. Determination of Measurement Method for Each Evaluation Indicator

Pedestrian Fatality Risk

A lane shoulder is a road facility installed as a part of the traffic road to protect the structural part of the road and to allow visibility from vehicles and evacuation. To ensure the safety of pedestrians on a road without distinction between traffic and pedestrian roads, the lane shoulder width is sometimes increased. The width of the lane shoulder built on the right-hand side of the traffic road is defined at an equal or higher level based on the road type and design speed, as shown in

Table 1. Criteria for designing the lane shoulder width [25].

Road Type			Minimum Lane Shoulder Width (m)
			Rural Area	Urban Area	Compact Car Lane
Highway			3.00	2.00	2.00
Road	Design speed (km/h)	≥80	2.00	1.50	1.00
		60–80	1.50	1.00	0.75
		<60	1.00	0.75	0.75

.

The lane shoulder width may constantly vary within the short distance according to land use and other factors. Therefore, suggesting a representative value for the given distance is challenging. Using the minimum or maximum lane shoulder width at the candidate site would bias the score of the evaluation indicator for the site. Thus, the weighted average reflecting the road length should be used to score the lane shoulder width.

Numerous studies have suggested the possible relationship between speed and traffic accidents. Rosén and Sander [26] reported the relationship between vehicle running speed and pedestrian mortality; pedestrian mortality decreased by approximately 25% from 20% to 15% when the vehicle running speed decreased from 60 km per h to 50 km per h, while the mortality was markedly reduced to below 10% when the speed was 30 km per h. Like other countries, the presence or absence of speed enforcement cameras in Korea affects the driving speed, but this study does not reflect the fact that speed enforcement cameras usually reduce speed only at points other than sections. Thus, for the speed limit, the one at the candidate site is identified and scored based on the criteria of 60 km per h.

For traffic volume at the candidate site, official data should be applied. Typically, AADT, the mean of 24 h traffic volume recorded for 365 days at a single target area, is used. A high traffic volume at a single site implies a high probability of vehicle-to-pedestrian collision, and in the absence of a pedestrian road at a site with a high AADT, the pedestrian fatality risk would be high. Thus, the AADT at each site, as the data published by the government [27], is used in this study as the evaluation indicator of traffic volume. As the adequate traffic volume for a two-lane national highway of the LOS C is 5300 vehicles, the traffic volume is scored based on the criteria of 5000 vehicles [28].

A traffic accident is a crucial and objective attribute of the fatality risk at a given candidate site, and the data collection is convenient as the government presents the traffic accident statistics for all road sites each year. Thus, the traffic accident in this study is scored based on the criteria of 15 cases in the past three years at a given site, as shown in

Table 2. Estimation and score of pedestrian fatality risk.

Evaluation Indicator			Grade (Score)
Attribute	Quantitative/Qualitative	Evaluation Criteria	A (1.0)	B (0.8)	C (0.6)	D (0.4)	E (0.2)
Lane shoulder width	Quantitative	Mean weight of lane shoulder width at the candidate site (m)	≤0.5	≤1.0	≤2.0	≤3.0	>3.0
speed limit	Quantitative	Speed limit at the candidate site (km/h)	≥80	≥70	≥60	≥50	<50
AADT	Quantitative	AADT at the candidate site	≥20,000	≥10,000	≥5000	≥2500	<2500
Traffic accident	Quantitative	Traffic accidents at the candidate site in the past three years (case)	≥50	≥30	≥15	≥5	<5

Note. AADT, average annual daily traffic.

.

Surrounding Infrastructure

The presence of a pedestrian road close to the candidate site for pedestrian road construction has two implications. First, there is already a significant pedestrian demand in the area close to the candidate site. Second, if a new pedestrian road is constructed at the candidate site to be connected with the existing road, the connectivity for pedestrian travel can be improved. Thus, the connectivity to the facilities at the candidate site and the existing pedestrian road was scored in this study. Additionally, in the case a direct connection with the existing road is prevented, the potential connectivity to a pedestrian road within walking distance could be scored qualitatively. The residents in a village close to a national highway in a rural area tend to walk for accessibility to the key facilities within one km of distance. Specifically, the main pedestrian-generating facilities in a rural area, such as the village office and community hall, are critical factors in determining the site for pedestrian road construction. Thus, the presence or absence and the number of public facilities, including post offices, educational facilities, and community halls, were applied. Additionally, to apply the size of the key facilities and their distance from the candidate site, whether the key facilities were located in a direct impact zone (within 100 m) or an indirect impact zone (within 500 m) was qualitatively evaluated.

Pedestrian travel can be divided into the trip to access key facilities and public transportation for long-distance travel. Notably, the connectivity to a public transportation facility, such as a bus station and pedestrian road, is critical to pedestrian demand and safety. Thus, the connectivity between the candidate site and the public transportation facility was applied in this study. Additionally, as with the key facilities, impact zones were set for the qualitative evaluation of the distance between the candidate site and the public transportation facility. For national highways in rural areas, facilities related to walking, such as rivers, parks, and tourist sites, are found in the vicinity. Such facilities generate pedestrian travel and are, therefore, an influencing factor in pedestrian travel and safety. Thus, the size of tourist sites and their distance from the candidate site were included by the evaluation indicator in this study, and impact zones were set for the qualitative evaluation.

The evaluation indicator and criteria on surrounding infrastructure include both quantitative and qualitative measures to ensure that the evaluator’s judgment for field investigation is critical. Additionally, the scoring must be based on the qualitative criteria that define the grade of the site of interest within the entire area of the candidate site. Thus, for scoring, five grades were set by 20% intervals for the scoring of each attribute, as shown in

Table 3. Estimation and score of surrounding infrastructure.

Evaluation Indicator			Grade (Score)
Attribute	Quantitative/Qualitative	Evaluation Criteria	A (1.0)	B (0.8)	C (0.6)	D (0.4)	E (0.2)
Connectivity to pedestrian roads	Quantitative + Qualitative	Connectivity to pedestrian roads within the area of candidate site (0–1)	very high (upper 20%)	high (20– 40%)	moderate (40– 60%)	low (60– 80%)	very low (80– 100%)
Connectivity to key facilities	Quantitative + Qualitative	Connectivity to public facilities and other key facilities (community hall, village office, etc.) (0–1)
Connectivity to traffic facilities	Quantitative + Qualitative	Connectivity to public transportation facilities (bus station, train station, car park, etc.) (0–1)
Tourism	Quantitative + Qualitative	Connectivity to rivers, parks, and historic sites (0–1)

.

Potential Pedestrian Demand

The population of an area, including the candidate site, is the most common indicator of transportation demand and a primary factor in the prediction of traffic road demand. Thus, the population of the minimum administrative district at the candidate site was defined as the attribute of pedestrian population. However, the population could not be applied directly as the rate of pedestrians was higher for individuals aged ≤20 s and ≥50 s than those aged between 20–40 s who use vehicles due to active economic activities. Hence, the rate of pedestrians was adjusted using the distribution rate of public transportation use in each age group, as suggested in the survey of the current use of public transportation and the current population data [29].

Many commercial facilities within walking distance or the candidate site indicate a high pedestrian demand. Thus, the density of pedestrian-generating facilities was selected as an evaluation indicator. The density was estimated based on the number of pedestrian-generating facilities against the road length and the candidate site, while the impact zone was set to an area within 500 m distance from the start and end and the sides of the candidate site as shown in Figure 5. The red line means the pedestrian road to be installed. The criteria for pedestrian population and pedestrian-generating facility can be modified according to the road type, and in this study, the scoring was based on the pedestrian population of n = 750 and pedestrian-generating facilities of n = 10, as shown in

Table 4. Estimation and score of potential pedestrian demand.

Evaluation Indicator								Grade (Score)
Attribute	Quantitative/Qualitative	Evaluation Criteria						A (1.0)	B (0.8)	C (0.6)	D (0.4)	E (0.2)
Pedestrian population	Quantitative	Permitted number of pedestrians reflecting age; dependent rate of pedestrian travel						≥2000	≥1000	≥750	≥500	<500
		Category	<20 years	<30 years	<40 years	<50 years	≥50 years
		Pedestrian percentage (%)	69.87	25.84	28.64	26.21	41.01
		Quantitative measure (%)	36	13	15	14	21
Pedestrian-generating facility	Quantitative	Number of pedestrian-generating facilities per road length (Number of facilities/road length (km) at the candidate site)						≥30	≥15	≥10	≥5	<5

.

Road Manager’s Opinion

Compiling the road manager’s opinions is a quantitative indicator of the need for the prioritization of the candidate site for pedestrian road construction, requested by the respective road management agency. First, the road manager sets the priority across the candidate sites from that most requiring the construction and scores the sites accordingly. To illustrate, if the road manager sets the priority across 29 sites from the 1st to the 29th, the score of the 1st rank site is 100 (29/29*100), and that of the 29th rank site is 3.45 (1/29*100).

$$\mathrm{Score}=\frac{R}{S}\times 100$$

(1)

where R denotes the rank in the site, S denotes the number of site. The construction rationale is quantified through keyword analysis based on the one suggested by the road manager. The main keywords supporting the need for pedestrian road construction are civil complaints, connectivity to public transportation facilities, fatal accidents, and traffic volume and speed. The scoring in quantification is based on the inclusion of the main keyword, as shown in

Table 5. Estimation and score of road manager’s opinion.

Evaluation Indicator			Grade (Score)
Attribute	Quantitative/Qualitative	Evaluation Criteria	A (1.0)	B (0.8)	C (0.6)	D (0.4)	E (0.2)
Manager priority	Quantitative	Scoring the road manager priority	very high (1.0–0.8)	high (0.8–0.6)	moderate (0.6–0.4)	low (0.4–0.2)	very low (0.2–0.0)
Construction rationale	Quantitative	Analyzing the keywords of pedestrian road construction rationale	n = 4	n = 3	n = 2	n = 1	n = 0

.

Weighting of Each Evaluation Indicator

The evaluation indicators in the priority setting for the pedestrian road construction should be weighted according to the road type. The evaluator may determine the weightage of an evaluation indicator based on the characteristics of the road and the candidate site. This study applied the Analytic Hierarchy Process (AHP) to estimate the weight of each evaluation indicator. The AHP is a technique in multi-criteria decision making whereby the influencing factors of nonlinearity are selected, and their hierarchy is composed of the pairwise comparison of the importance of each indicator and its weightage and set priority [30]. It is a widely used qualitative method in multi-criteria decision making. The advantage of AHP is that it illustrates how possible changes in priority at upper levels have an effect on the priority of criteria at lower levels. However, the disadvantage of AHP is the complexity of this method which makes its implementation quite inconvenient [25].

For the survey participants, those with adequate knowledge of pedestrian roads and those capable of evaluating a given project from the perspective of public benefits must be recruited, as the construction of pedestrian roads is a public project. In this study, the AHP survey was conducted on 20 road traffic specialists with expertise on pedestrian roads over 20 years to estimate the weight of each suggested evaluation indicator. Although the appropriate number of AHP surveyors depends on the research field, the experience of the relevant field is more important than the number of surveyors, considering the importance of the survey results for the comparison of each indicator. Therefore, in this study, specialists with experience on pedestrian roads for more than 20 years were selected.

This study carried out pairwise comparisons to compute the weight of each evaluation indicators based on the results of the questionnaire on importance of each evaluation indicators. Pairwise comparisons is a technique to quantify the relative importance (weight) between the components based on the survey. It applies a mathematical technique to the binary comparison matrix so that decision-making elements can be compared by the hierarchy. Element of the matrix, aij, computes how much the element i is preferable to the element j for the n elements. Then, the comparison matrix A = (a_ij); ij = 1, 2, …, n can be obtained as Equation (2):

$$\mathrm{A}=\left[\begin{array}{ccc}1& \cdots & a_{1n}\\ ⋮& 1& ⋮\\ a_{n1}& \cdots & 1\end{array}\right]$$

(2)

When each weight between n number of elements is expressed as w_i (i = 1, 2, …, n), the element, a_ij, of a comparison matrix A, represents weight rate between row coordinate i and column coordinate j. Equation (3) shows the mathematical expression of an element, a_ij, of a comparison matrix A:

$$a_{ij}=\frac{w_i}{w_j} \left(\mathrm{for} i,j=1,2,\cdots ,n\right)$$

(3)

Using Equation (3), pairwise comparison matrix of Equation (2) can be expressed as Equation (4):

$$\mathrm{A}=\left[\begin{array}{ccc}\frac{w_1}{w_1}& \cdots & \frac{w_1}{w_n}\\ ⋮& \ddots & ⋮\\ \frac{w_n}{w_1}& \cdots & \frac{w_n}{w_n}\end{array}\right]$$

(4)

The correctness of the conducted paired comparison assessment of the criteria indicated in the analysis is verified by calculating, for each matrix A, the consistency index CI.

$$\mathrm{CI}=\frac{{\lambda }_{max}-n}{n-1}$$

(5)

where: n—matrix dimension, λmax—maximum matrix eigenvalue.

To check if paired comparison assessment of the indicators is consistent, Saaty [25] proposed what is called the Consistency Ratio (CR), which is a comparison between Consistency Index and Random Consistency Index (RI) as shown in

Table 6. Random index.

n	1	2	3	4	5	6	7	8	9	10
RI	0	0	0.52	0.89	1.11	1.25	1.35	1.40	1.45	1.49

. If the value of Consistency Ratio is smaller or equal to 10%, the inconsistency is acceptable.

$$\mathrm{CR}=\frac{\mathrm{CI}}{\mathrm{RI}}$$

(6)

As a result of the analysis, the evaluation indicators for both the first hierarchy and second hierarchy were calculated to be below 0.1. Therefore, the pairwise comparison is acceptable and score weight can be useful in decision making.

The final score of the candidate site was estimated by multiplying the score of each attribute based on the criteria by the estimated weight of the attribute, as shown in

Table 7. Weighting of each hierarchy for evaluation indicators.

Evaluation Indicator in the First Hierarchy	Weight of the First Hierarchy	Evaluation Indicator in the Second Hierarchy	Weight of the Second Hierarchy
Pedestrian fatality risk	0.443	Mean lane shoulder width	0.046
		Speed limit	0.127
		AADT	0.086
		Traffic accident	0.185
Surrounding infrastructure	0.312	Connectivity to pedestrian roads	0.127
		Connectivity to key facilities	0.061
		Connectivity to traffic facilities	0.071
		Tourism	0.053
Potential pedestrian demand	0.136	Pedestrian population	0.06
		Density of pedestrian-generating facilities	0.075
Manager’s opinion	0.109	Manager priority	0.024
		Construction rationale	0.085

Note. AADT, average annual daily traffic.

. From the AHP survey result, the weight of the first hierarchy was as follows: 0.443 for pedestrian fatality risk, 0.312 for surrounding infrastructure, 0.136 for potential pedestrian demand, and 0.109 for manager’s opinion, in the order of importance; the weight of the second hierarchy was the highest at 0.418 for traffic accident among the attributes of pedestrian fatality risk, 0.408 for connectivity to pedestrian road among the attributes of surrounding infrastructure, 0.556 for the density of pedestrian-generating facilities among the attributes of potential pedestrian demand, and 0.778 for the construction rationale from manager’s opinion.

4. A Case Study

Random sampling has been conducted to apply the evaluation indicators and their individual estimation methods and weights for the pedestrian road priority setting on the national highway as presented in this study to an actual area.

A case study was conducted to apply the evaluation indicators and their individual estimation methods and weights for the pedestrian road priority setting on the national highway as presented in this study to an actual area. The goal was to verify the applicability of the suggested evaluation indicators and estimation methods in practice. With the cooperation of the road management agency, the data were collected for 385 candidate sites requiring pedestrian roads on national highways. The candidate sites in the case study were located along a 0.4 km length for which the agency had requested the evaluation to verify the need for pedestrian road construction. The sites were on the road close to a residential area comprising an apartment complex and other facilities, with the main pedestrian travel characterized by the connectivity to the bus station as shown in Figure 6.

First, the evaluated score of pedestrian fatality risk was 23.8 as shown in

Table 8. An example of applying the attribute of pedestrian fatality risk.

Attribute	Survey Data	AHP Weight (1) (ⓐ)	AHP Weight (2) (ⓑ)	Score (ⓒ)	Converted Score (ⓐ × ⓑ × ⓒ)
Lane shoulder width	0.625 m	0.443	0.102	0.8	0.036
Speed limit	60 km/h	0.443	0.287	0.6	0.076
AADT	8954	0.443	0.193	0.6	0.051
Traffic accident	9 cases	0.443	0.418	0.4	0.074
Total				244th rank	0.238 (23.8)

Note. AADT, average annual daily traffic; AHP, Analytic Hierarchy Process.

, placing the candidate site at 244th priority across 385 sites. This may be due to the low traffic volume and number of traffic accidents compared to other sites.

Subsequently, the evaluated score of surrounding infrastructure was 25.5, with the second rank priority, as shown in

Table 9. An example of applying the attribute of surrounding infrastructure.

Attribute	Survey Data	AHP Weight (1) (ⓐ)	AHP Weight (2) (ⓑ)	Score (ⓒ)	Converted Score (ⓐ × ⓑ × ⓒ)
Connectivity to pedestrian roads	18	0.312	0.408	0.90	0.115
Connectivity to key facilities	16	0.312	0.193	0.80	0.048
Connectivity to traffic facilities	17	0.312	0.228	0.85	0.060
Tourism	12	0.312	0.168	0.6	0.031
Total				2nd rank	0.255 (25.5)

Note. AHP, Analytic Hierarchy Process.

. This may be due to the existing pedestrian road at the site entailing an urgent need for construction considering connectivity and proximity to public transportation facilities such as the bus station.

The evaluated score of potential pedestrian demand was 12.0, with the 42nd rank priority as shown in

Table 10. An example of applying the attribute of potential pedestrian demand.

Attribute	Survey Data	AHP Weight (1) (ⓐ)	AHP Weight (2) (ⓑ)	Score (ⓒ)	Converted Score (ⓐ × ⓑ × ⓒ)
Pedestrian population	3177	0.135	0.444	1.0	0.060
Pedestrian-generating facility	17.5	0.135	0.556	0.8	0.060
Total				42nd rank	0.120 (12.0)

Note. AHP, Analytic Hierarchy Process.

. This may be due to the high score on pedestrian population compared to other sites and the high density of pedestrian-generating facilities. Additionally, the evaluated score of manager’s opinion was 10.4, with the fourth rank priority as shown in

Table 11. An example of applying the attribute of manager’s opinion.

Attribute	Survey Data	AHP Weight (1) (ⓐ)	AHP Weight (2) (ⓑ)	Score (ⓒ)	Converted Score (ⓐ × ⓑ × ⓒ)
Management agency priority	0.8	0.109	0.222	0.8	0.019
Road construction rationale	n = 4	0.109	0.778	1.0	0.085
Total				4th rank	0.104 (10.4)

Note. AHP, Analytic Hierarchy Process.

.

The final score result showed that the candidate site in the trial application ranked 41st across 385 sites with a total score of 71.6 based on the evaluated score of 23.8 for pedestrian fatality risk (244th/385 sites), the score of 25.5 for surrounding infrastructure (2nd/385 sites), the score of 12.0 for potential pedestrian demand (42nd/385 sites), and the score of 10.4 for manager’s opinion (4th/385 sites). The site on which the case study was conducted had markedly high scores for connectivity and manager priority but a low score for pedestrian fatality risk with a relatively high level of AHP weight compared to other sites.

The evaluation methods tested through the trial application were applied to all 385 sites, and the scores and ranks of each indicator from the 1st to the 385th are presented in

Table 12. Score and rank of each evaluation indicator at each candidate site (n = 385).

Site	Score of Pedestrian Fatality Risk	Score of Surrounding Infrastructure	Score of Potential Pedestrian Demand	Score of Manager’s Opinion	Sum
1	42.4	24.52	12.0	8.95	87.87
2	42.4	22.16	12.3	10.9	87.76
3	43.3	16.04	13.5	10.82	83.65
4	40.76	23.68	12.0	7.13	83.57
5	41.6	22.75	13.5	5.32	83.17
6	39.6	19.31	12.0	10.05	80.96
7	43.3	21.38	10.5	5.25	80.43
8	35.72	21.21	13.5	7.50	77.93
9	39.06	18.4	13.5	6.42	77.38
10	41.6	13.86	13.5	8.4	77.36
11	44.2	14.5	12.3	5.8	76.8
12	41.66	12.6	13.5	8.9	76.66
· · ·	· · ·	· · ·	· · ·	· · ·	· · ·
385	21.12	9.02	2.7	1.8	34.64

. The scoring was based on the AHP weight assigned by an expert and the result of the field investigation. The proposed methodology of priority setting for the construction of pedestrian roads is deemed applicable to other road types than national highways, such as provincial roads.

5. Conclusions

This study aimed to propose a methodology of priority setting for the construction of pedestrian roads using a set of evaluation steps distinguished from the conventional economic feasibility analysis (B/C) and a methodology applicable to national highways. Four evaluation indicators were selected in this study, including pedestrian fatality risk, surrounding infrastructure, potential pedestrian demand, and manager’s opinion, with a total of 12 attributes across the indicators. Each attribute was designed to allow a quantitative or qualitative evaluation, with a method of estimating and scoring each attribute according to set criteria. In the weighting of evaluation indicators, the AHP was applied as the most widely used method in multi-criteria decision-making. The highest weight was placed on pedestrian fatality risk and surrounding infrastructure among the evaluation indicators and on traffic accidents of pedestrian fatality risk among the attributes.

To determine the practical application of the proposed methodology, an evaluation of a single site was conducted first, and in the same way, 385 sites in total were assessed for priority setting. The evaluation addressed 12 attributes of four indicators, and the total score range was 87.87–34.64.

The four indicators and attributes presented in this study were composed of independent indicators and attributes, but some attributes may be related. Therefore, the indicators and attributes presented in this study can be changed or added/deleted in other studies. In addition, the method of classifying grades for each evaluation indicators was to follow national standards or reference such as shoulder width or speed limit. On the other hand, evaluation indicators without national standards or reference were classified using the turning point or the same ratio of the data used in the case study of this study. As for the objective criteria for classifying the grade of the evaluation indicators, further research is needed in the future.

The methodology of priority setting for pedestrian roads, as proposed in this study, can be applied to the priority setting for pedestrian roads and also smaller-scale SOC projects, including bicycle roads. It is also a reliable alternative to resolve the drawbacks of the plans with a top-down approach that cannot reflect on-site opinions. Notably, while priority setting requires relatively high levels of time and effort due to the diversity of evaluation indicators and their attributes, the methodology is applicable for efficient budget allocation. Furthermore, if an imaging device such as MMS (Mobile Mapping System) could be used in the field investigation rather than a human survey, the time taken by evaluation is likely to be reduced.