1 Introduction and Literature Review

The level of service (LOS) term was first introduced in 1965 for highways, as a version of the Highway Capacity Manual (HCM) [1]. In the 1965 version of the manual, the LOS was divided into six classes defined with letters from A (best) to F (unacceptable), each of which represented a range of operating conditions with definitions based on a combination of travel time and the ratio of the traffic flow rate to the road capacity. At the same time, researchers developed LOS standards for pedestrians. These standards were classified into various LOS values, again ranging from A to F, with LOS A representing the threshold of unimpeded free flow and LOS F at the critical density or breakdown of movement continuity. These standards contribute mainly to the designs of engineers and architects.

Fruin researched pedestrians in the early 1970s. “Pedestrian Planning and Design” [2] has been cited in many of the current guidelines for pedestrian planning. This research has become the standard for many subsequent building design and planning operations.

Since its first introduction by Fruin [3], LOS has been used extensively to explain how pedestrians feel in various walking conditions (walking/sitting/queuing). This method is very useful for describing the impacts of pedestrian volume increases on the pedestrian LOS. The general objective of determining LOS standards for pedestrians includes many potential research areas that could not be included within the limitations of a single project [4]. In “Pedestrian Planning and Design”, Fruin notes a number of attributes that facilitate the LOS term and addresses the fundamental space requirements of a single pedestrian. This concept was redefined in relation to several traffic conditions in the 1985 version of the HCM [5]. Therefore, Fruin agrees with the HCM that pedestrian movement is very important. Also, Goffman [6] states, “pedestrians can twist, duck, bend, and turn sharply, and therefore, unlike motorists, they can safely count on being able to… before impending impact”. Whereas vehicle density is important for explaining vehicle flow and its effects on motorists’ perceptions of the flow, a quite different approach is required to explain the comfort needs of pedestrians.

A new LOS called “Pedestrian LOS” (PLOS) is under development for pedestrians. PLOS models are believed to determine reasonable LOS values for pedestrians. Walking speed, pedestrian spacing, and probability of conflict at various traffic concentrations are the main factors that determine each service level. These quantities are linked directly to factors that affect comfort and safety, so these are reflecting pedestrian perception of the degree to which a facility is “pedestrian friendly”. Seneviratne and Morrall [7] considered the perceived quality of service for ranking and designing pedestrian facilities in which the characteristics of trip makers, that is, the trips and physical features, are effective. Sarkar [8] proposed six service levels on the basis of the quality of a facility that is provided for pedestrians (Fig. 1).

Fig. 1
figure 1

Determining factors affecting quality of spatial service level [8]

Full size image

Sarkar [8] attempted to consider the influences of environmental factors on the PLOS and put forward numerous qualitative and quantitative measures that included almost all of the possible factors and proposed some corresponding PLOS models [9].

Some studies have reported that PLOS values and models include changeable factors and are related to some specifications. Mori and Tsukaguchi [10] conducted a study of the design and evaluation of Asian pedestrian facilities and proposed a new method for the evaluation of LOS values under different flow conditions. Some researchers proposed LOS standards for evaluating pedestrian facilities in Manila according to the behavioral characteristics of pedestrians and the preferences developed for factors affecting pedestrian route choices [11]. Khisty [12] found that qualitative environmental factors are as important as quantitative PLOS factors, such as flow, speed, and density in the planning, design, and evaluation of pedestrian facilities.

Henson [13] summarized research studies on the PLOS for pedestrian facilities related to signalized crosswalks and underground stations. Research related to the PLOS standards for signalized crosswalks in Hong Kong was conducted by Lam and Lee [14]. The PLOS measures presented in the HCM [15] for signalized crosswalks describe the amount of congestion (quality of flow) faced by pedestrians and employ several LOS measures for walkways and crosswalks. Also, for pedestrians in Hong Kong, time was found to be the most important factor of concern, with interest in the environment being a low priority. Lee et al. [16] redefined the LOS boundaries for signalized crosswalks in Hong Kong commercial areas in terms of pedestrian density, pedestrian flow rate, and walking speed through a study of stairways in Hong Kong underground stations. Also, Erkan and Hastemoglu [17] developed a new PLOS standard for a subway station in Turkey and provided a link between PLOS and pedestrian speed.

PLOS models are utility models for urban areas and are used to facilitate the design of urban streets. Some studies have proposed other design factors that could be used to evaluate the PLOS in urban areas. These included the path width, road surface quality, obstructions, crossing opportunities, connectivity, sidewalk paths, vehicle conflict, pedestrian volume, mix of path users, and security. The developed PLOS was based on a pedestrian opinion survey. Hostovsky et al. [18] examined freeway service quality for two different road user groups: urban commuters and rural commuters. Urban pedestrians were found to be more concerned about travel time, whereas rural commuters were more concerned about maneuverability. Currently, pedestrians in the US are not permitted on freeways, and it is unlikely that rural pedestrians would be concerned with PLOS. For this reason, the use of different density thresholds for urban and rural freeways was suggested [19].

Researchers have also studied street attractiveness by considering design developments (e.g., roadside restaurants, malls, and street furniture). On the basis of the increase in roadside development and the decrease in pedestrian flow quality (in terms of pedestrian speed, density, etc.), later studies devised a qualitative analysis of pedestrians, which included new factors such as pedestrian safety, security, continuity, comfort, and convenience.

Conversely, researchers have evaluated the relative importance of qualitative factors, such as comfort, safety, security, attractiveness, convenience, and accessibility to measure the LOS of pedestrians [20].

Muraleetharan et al. [21] defined the “overall LOS” as a single index that combines the effects of different factors and underlines the importance of considering the total utility value when defining the PLOS. The combined effect of the factors affecting the PLOS was achieved through a unified technique.

LOS studies have also been performed with the help of mathematical models that have considered the following factors: availability of sidewalk, lateral separation between pedestrians and vehicular traffic, traffic volume, and vehicle speed [22–25].

Further, researchers evaluated PLOS in terms of pedestrian space as well as evasive movements. The results indicated that the evasive pedestrian movements could better explain the pedestrian-perceived LOS [26].

Measurement criteria are necessary to measure the overall operating conditions that depend specifically on pedestrian needs, including route, path, facility, and space quality. A spatial design that considers user needs must include physical factors that influence the comfort, satisfaction, and safety and security of pedestrians.

Some buildings are developed for a certain and constant purpose that do not allow for any alternative utilization [27]. In particular, the design and management issues related to transportation buildings are considered major concerns and are scrutinized with special importance [28, 29]. These buildings serve various groups of people who have different ethnic, cultural, economic, and social characteristics. Different group characteristics result in different demands for both socio-spatial and satisfaction needs. Human characteristics and behaviors are directly related to human scale, which is fundamental to the LOS. Therefore, a comprehensive design approach must take human scale into account, assuming it aims to achieve successful LOS arrangement and application. It is necessary to standardize LOS calculations to meet the demands of different users.

Most LOS research accepts the former LOS standards and produces some inferences. However, when the previous studies are evaluated, it becomes evident that there are very different LOS standards in use for very similar research. This situation has resulted in the production of different LOS data for the same space.

2 Problem Statement

When previous research is examined, it is obvious that there is no consensus on some critical issues, even in recent studies. In that sense, the HCM data, which were used to develop the LOS concept based on that of vehicle density, and Fruin’s data, which defined the LOS in an architectural space by linking LOS values with pedestrians, produced different results and values. The main reason for the variations in the LOS results of different studies can be linked to the LOS approach, which is defined as the number of humans calculated per unit area. Thus, an LOS value is directly related to the anthropometric scale, and different researchers adopt different body dimensions and comfort scales. Figure 2 demonstrates the different anthropometric body scales of Fruin and the HCM.

Fig. 2
figure 2

Body scales of Fruin and HCM [3, 32]

Full size image

Anthropometry is a science that deals with the measurement of the size, weight, and proportions of the human body [30]. Specifically, the methods of measuring the dimensions of the shoulders, muscles, and so on, of the human body are studied [31]. Anthropometrics measure the dimensions of the body and other physical measurements so that body measurements can be used to describe the sizes, shapes (forms or structures), and compositions of individuals.

Fruin [3] illustrates the importance of body dimensions and describes the differences as follows: “Body depth and shoulder breadth are the primary human measurements used in considering pedestrian spaces and facilities. Shoulder breadth is a major factor in the design of doorways, stairways, and so on”.

On the other hand, the HCM study [32] described a pedestrian ellipse zone and stated that this zone plays an important role in shaping the space. Because body dimensions are so important in space shaping, each proposed value is important as well. In that sense, the works [33–35] on anthropometric studies were compiled, and the newly proposed body dimensions are specified in Table 1.

Table 1 Anthropometric sizes of different world populations
Full size table

The anthropometric sizes were obtained from studies conducted on various body dimensions, cover a large profile of the world’s population, and indicate an “average-sized” individual of “45.12 cm by 24.40 cm”. It is completely understandable that the existing LOS values are different from each other due to the considerable inconsistencies in human body dimensions that were used to determine LOS standards. The other studies conducted based on the LOS concept, including the works of Fruin and the HCM, are summarized in Table 2.

Table 2 LOS standards developed by different research
Full size table

Today, many institutions in various countries have been developing analyses based on different LOS factors. It has been claimed that it is not necessary to develop universal agreement on LOS standards, but it is beneficial and necessary to develop a LOS determination method. This necessity has been linked to recent studies, which assert that different human communities have different anthropometric sizes. The primary human body size dimensions that influence the anthropometric scales are the “breadth” and “depth” values which are used in most studies.

3 Research Method

A software program called Laborer Image Analysis Software (LIAS) was developed in this study (see Fig. 3).

Fig. 3
figure 3

Partial interface of Laborer Image Analysis Software (LIAS)

Full size image

The program transfers a plan image of a human body to a body-covering area, which is also calculated by the software through an image analysis method. The Fruin and HCM values can be compared with the help of this software. The program can also calculate the anthropometric values (breadth and depth) that have been used in recent studies. It provides mean values and reveals the major differences from the previous study results. The main reason that the proposed LOS values take the Fruin and HCM values into account is the availability of and access to clear explanations of the anthropometric scales used in those studies.

The LIAS program was used to analyze Fruin’s body dimension proposal and determined the total covered area of a single body to be 0.15 m2, whereas it calculated the same quantity to be 0.24 m2 based on the HCM values. When the proposed anthropometric scales are considered, the total number of pedestrians can be calculated from the m2/pedestrian formula as:

$$\Rightarrow \frac{{3.2 \;{\text{m}}^{2} / {\text{ped}}}}{{0.15 \;{\text{m}}^{ 2} }} = 21.33{\text{ pedestrians}}$$

According to the proposed anthropometric values, the covered area of a single body was analyzed through LIAS and accepted as 0.10 m2. This accepted value was used in the following equation to find the LOS A value: 21.33 pedestrians × 0.10 m2 = 2.13. This result (2.13) represents the LOS A value. However, these values are only valid if the Fruin anthropometric values are taken into consideration. Calculation with the same method for the HCM anthropometric values was also necessary. In this case, the equation becomes: (m2/pedestrian) = 12/0.24 = 50 pedestrians. According to the proposed anthropometric values, the covered area of a single body was analyzed through LIAS and accepted to be 0.10 m2. This accepted value was used in the following equation to find the LOS A value: 50 pedestrians × 0.10 m2 = 5. This result (5) represents the LOS A value.

It is inevitable that each value will show a different result when it is calculated using different anthropometric values. The analysis and comparison of all of the LOS values, including the proposed anthropometric data, are given in Table 3.

Table 3 Proposed LOS values derived from Fruin and HCM values
Full size table

As shown in Table 3, two LOS values were proposed for use rather than the existing LOS values.

At the same time, A1 is the proposed LOS(1) and proposed LOS(2) for LOS A; B1 is the proposed LOS(1) and proposed LOS(2) for LOS B; C1 is the proposed LOS(1) and proposed LOS(2) for LOS C; D1 is the proposed LOS(1) and proposed LOS(2) for LOS D; E1 is the proposed LOS(1) and proposed LOS(2) for LOS E; F1 is the proposed LOS(1) and proposed LOS(2) for LOS F. We now present the attempt to create an intersection of sets to predefine the LOS values. An intersection of sets is defined as a grouping of the common elements of two or more sets (given two sets A and B, the intersection A ∩ B is the set that contains elements or objects that belong to both A and B. Basically, A ∩ B can be found by identifying all of the elements that A and B have in common).

To try to be unified with an intersection set and if the “x” value is developing LOS value.

  • For the LOS A value: A1 = {x|x Є R, x > 2.13} and Az = {x|x Є R, x > 5}

    A1 ∩ Az = {x|x Є R, x > 5}

  • For the LOS B value: B1 = {x|x Є R, 1.51 < x < 2.13} and Bz = {x|x Є R, 1.08 < x < 5}

    B1 ∩ Bz = {|x Є R, 1.08 < x < 2.13}

  • For the LOS C value: C1 = {x|x Є R, 0.93 < x < 1.51} and Cz = {x|x Є R, 0.64 < x < 1.08}

    C1 ∩ Cz = {x|x Є R, 0.93 < x < 1.08}

  • For the LOS D value: D1 = {x|x Є R, 0.60 < x < 0.93} and Dz = {x|x Є R, 0.41 < x < 0.64}

    D1 ∩ Dz = {x|x Є R, 0.60 < x < 0.64}

  • For the LOS E value: E1 = {x|x Є R, 0.33 < x < 0.60} and Ez = {x|x Є R, 0.22 < x < 0.41}

    E1 ∩ Ez = {x|x Є R, 0.33 < x < 0.41}

  • For the LOS F value: F1 = {x|x Є R, x < 0.33} and Fz = {x|x Є R, x < 0.22}

    F1 ∩ Fz = {x|x Є R, x < 0.33}

On the other hand, the intersection set developed for the Fruin and HCM values reveals that, of the intersection points, the beginning and end values are not suitable as LOS standards (Table 4).

Table 4 Reference values are inconsistent according to intersection set of both values: (m2/pedestrian)
Full size table

In an LOS system, the values need to follow each other, because it is not possible to adopt standard values for the intermediate values that fall outside of the reference values. Due to the fact that an intersection set cannot develop predefined reference values, it is necessary to reanalyze the values developed in Table 2 using a different method.

This different analysis method is called “emptiness area analysis”. In other words, due to the variability of human anthropometric values, it is necessary in LOS studies to calculate a unit area from the empty space around the body surface area.

Emptiness area defines a spatial area given by the body surface area subtracted from the total area (emptiness area = total area − body surface area).

If “the emptiness area” is defined as “TER”, formula (1) can be written. In this study, the emptiness area was determined on the basis of the developed formula:

$$\begin{aligned} {\text{TER}} = & \frac{{{\text{LOS}}\left( { \hbox{min} } \right){\text{X}} - {\text{BSA}}}}{{{\text{LOS}}\left( { \hbox{min} } \right){\text{X}}}} \\ = & \frac{{{\text{LOS}} \left( { \hbox{min} } \right){\text{X}}}}{{{\text{LOS}} \left( { \hbox{min} } \right){\text{X}}}} - \frac{{\left( {\text{Body Surface Area}} \right)}}{{{\text{LOS}} \left( { \hbox{min} } \right){\text{X}}}} \\ & \quad {\text{TER = }}1 - \frac{{\left( {\text{Body Surface Area}} \right)}}{{{\text{LOS}}\left( { \hbox{min} } \right){\text{X}}}} \\ & \quad {\text{TER}} = 1 - \frac{{\left( {\text{Body Surface Area}} \right)}}{{{\text{LOS}}\left( { \hbox{min} } \right){\text{X}}}} \\ & {\text{TER}} + \frac{{\left( {\text{Body Surface Area}} \right)}}{{{\text{LOS}}\left( { \hbox{min} } \right){\text{X}}}} = 1 \\ \end{aligned}$$
(1)

In-formula annotations:

TER = The Emptiness Rate

BSA = Body Surface Area

LOS\(\left( { \hbox{min} } \right)x_{{}}\) = Minimum LOS value for specified service level

In addition, we defined the “emptiness area” as “TER” ⇒ P (A).

We also defined “fullness”, which is the opposite of “emptiness” ⇒ P(\(A^{\prime}\)).

The sum of the probabilities of all possible events equals 1. The rule of subtraction follows directly from this property. The probability that event A will occur is equal to 1 minus the probability that event A will not occur.

$$P\left( A \right) \, + \, P(A^{\prime}) = { 1} \Rightarrow {\text{TER}} + \frac{{\left( {\text{Body Surface Area}} \right)}}{{{\text{LOS}}\left( { \hbox{min} } \right){\text{X}}}} = 1$$
$$\begin{aligned} {\text{For example}},{\text{ in Fruin}},{\text{ LOS E}}: {\text{TER}} = & \frac{{{\text{LOS}}\left( { \hbox{min} } \right){\text{X}} - {\text{BSA}}}}{{{\text{LOS}}\left( { \hbox{min} } \right){\text{X}}}} \\ = & \frac{0.5 - 0.15}{0.5} \\ = & \, 0. 70 \, = \, P\left( A \right) \\ \end{aligned}$$
$$\frac{{\left( {\text{Body Surface Area}} \right)}}{{{\text{LOS}}\left( { \hbox{min} } \right){\text{X}}}} = \, P(A^{\prime}) = \frac{0.15}{0.5} = \, 0. 30$$
$$= \, P\left( A \right) \, + \, P(A^{\prime}) \, = \, 1 \Rightarrow 0.70 + \, 0.30 \, = \, 1$$

All of the LOS values that were examined confirmed that the formula is correct.

On the basis of recalculation using the emptiness area approach, the Fruin, HCM, and proposed values were developed, as shown in Fig. 4.

Fig. 4
figure 4

Comparison of newly proposed LOS values with existing LOS values for body scales

Full size image

4 Research Results

The emptiness area values obtained from the study were carefully examined. It is clear that emptiness areas belonging to different LOS levels have similar values. For instance, the LOS A value of Fruin and the LOS B value of Proposal (2) are almost the same.

This emptiness area problem has revealed the serious value inconsistency in the LOS system, particularly starting from the LOS C values. These inconsistencies are most prominent in the LOS D, E, and F values. Therefore, emptiness area calculation has become a very important criterion. In that regard, if two different male subjects (one from Hong Kong and the other from the US) who have different anthropometric sizes are compared, the importance of the emptiness area calculation can be better understood.

For the two aforementioned subjects, who belong to two different ethnic cultures, the body dimensions were determined to be 47 cm × 23.5 cm and 51.5 cm × 29 cm for the males from Hong Kong and the US, respectively. If both subject groups were located in the same space (dimensions: 1 m × 1 m = 1 m2), the comfort area could be calculated, as shown in Fig. 5.

Fig. 5
figure 5

Hong Kong and US male subjects located in the same area

Full size image

If two subject groups, each composed of four male subjects, were located in the same 1-m2 space, the Hong Kong subjects might feel more comfortable. Although this case may not seem problematic when the user numbers are relatively small, it becomes more complex when the number of users and space area increase. This complexity reveals the importance of emptiness area, which was the focus of this study.

5 Conclusion and Discussion

This study examined the LOS concept from a different perspective, which indicated the importance of human anthropometric dimensions, and developed software and a calculation method. Different cultures and communities have different anthropometric values, which result in different LOS analyses. Moreover, different researchers use these different values and obtain inconsistent results. The LOS standard is problematic because of these inconsistencies; thus, an LOS calculation method was proposed that includes a new emptiness area approach. It is believed that this new method will contribute to the understanding of human motion in general and to the spatial analysis and design of spaces for which LOS analysis is conducted.

Emptiness area is defined as the body surface area subtracted from the total area. Hence, emptiness area is a measurement that indicates pedestrian comfort and allows for suitable movement within the entire area. Furthermore, emptiness area is related to comfort area, which is very important for pedestrian satisfaction.

Thus, emptiness area analysis is important in the assessment of crowded places, particularly when the space is heavily used. LOS analysis is necessary for complex buildings to ensure user satisfaction and comfort. Different cultures with varying human body dimensions constantly use complex buildings, such as transportation structures, all around the world. In that sense, emptiness area and comfort area have become the focal points of LOS analysis and user satisfaction in today’s buildings, which accommodate greater varieties of people than ever before. This study developed user-friendly software, LIAS, which can be used to facilitate the work of designers and others in charge of developing and managing complex structures.

Future studies will focus on specific building types, such as train and subway stations, to test and enhance the validity of the software, as emptiness and comfort area need to be taken into account by the decision-makers who manage such complex buildings. Finally, anthropometric scales, which differ among communities, should also be considered by design teams and managers who are responsible for human satisfaction, comfort, and safety.