Analysis of Traffic Organisation in the Kiss-and-Fly Zone of Kraków Airport: Eye-Tracking Study

Abstract

When choosing the way to come to airports, quite a large number of passengers prefer when they are dropped off/picked up at airports using kiss-and-fly (K&F) zones. Such a travel option is associated with a special traffic organisation on the airport ground access. As there are no common regulations or standards when creating such a complex infrastructure object, it could be a challenge for drivers when searching and moving through it. Therefore, the main aim of the presented study was to assess and verify the eye-tracking technique as an objective tool, which can allow one to identify and estimate confusion points met by road users when using such an object. The field tests with 23 drivers were conducted in the K&F zone of Kraków Airport, and the data analysis focused on the traffic organisation and road signage as its key and integral parts. The eye-tracking approach allowed us to clearly find confusing situations for drivers as well as explain their reasons confirming its suitability and usefulness for the declared aim. Also, the perception of standardised and unstandardised signage of the K&F zone as well as the influence of route familiarity for drivers were discussed.

1. Introduction

Based on the geometry range within which transport infrastructure is developed, it can be divided into linear infrastructure elements such as roads and point infrastructure elements such as ports and airports, which could be also considered as complex infrastructure objects [1,2]. Although the second ones mainly act as connectors for the first ones, not only their appropriate planning and design but also the verification of their functionality for users are very important [3]. Moreover, some complex point infrastructure objects consist of few parts, which also often have a complex structure and consequently should be evaluated separately.

As a research tool in the field of transport, the first application of eye-tracking technique was in the 1970s [4,5]. Since that time, this technology strongly progressed, allowing us to cover both laboratory and field studies of different types of road users. In the first place, this method is widely used to understand drivers’ behaviour reacting to various visual stimuli [6]. Also, with the same aim eye-tracking technology, research was conducted with cyclists [7] and pedestrians [8]. A common feature of most such studies was a focus on searching general patterns, dependencies and relationships in the behaviour of the selected participants’ groups. There is quite a small amount of eye-tracking research works which have concentrated on such functional aspects of complex infrastructure objects or their parts as confusion points and problems met by road users when using them together and suggesting their further improvement. Here, the study at Kraków Główny railway station can be mentioned as an example [9].

Hence, the main purpose of this study was to assess and verify the eye-tracking technique with its standard metrics as an objective tool when searching and estimating confusions and problems of a complex infrastructure object/element faced by road users. The K&F zone of Kraków Airport was selected as a research object because of potential difficulties for drivers, which could be associated with the lack of accepted standards and regulations to design such an infrastructure element. The article is divided into six chapters. After the introduction, previous research works related to K&F zones themselves as well as the eye-tracking technique and its application to analyse road signage are discussed. Then, the organization and conduction of the field study including the description of selected area, equipment and participants are presented. Additionally, the approach to assess road signs is shown. Next, the outcomes concerning observations of road signs in the K&F zone are introduced in the form of an overview of road signs perception. Finally, each confusion point met by drivers in the K&F zone of Kraków Airport is discussed in detail. The article finishes with the conclusions.

2. Prior Related Works

2.1. Kiss-and-Fly Zone as a Research Object

As was mentioned above, the research object of the study was the K&F zone of Kraków Airport. Here, it is necessary to repeat that, in general, there are different options used by passengers to reach airports: car, drop-off, taxi or public transport [10]. Each of these options creates a different load of an airport ground (surface) access, which together with a terminal or terminal complexes are integrated parts of the landside area [11]. The K&F zone could be considered a special area assigned to drop off passengers at the airport (as well as pick them up) and as a variety of kiss-and-ride zone situated in airport surface access [12]; at the same time, trips associated with the usage of such zones are called K&F journeys [13]. In turn, the full definition a K&F zone (area) can be provided as follows: a complex infrastructure element, which is a part of airport ground access and situated in a certain location, where private cars/taxis are allowed to stop for a certain period of time (often up to 15 min) to pick up and drop off passengers and then leave without taking up space in the parking lot.

Despite the fact that a number of passengers preferring the ‘drop-off/pick-up’ mode of transport reaches sometimes even 43% [14], as well as the fact that K&F zones are quite common [15], there are not a lot of scientific articles related to these complex infrastructure elements; most of the research works concentrate on K&F journeys themselves and their negative environmental impact. The first broad discussion concerning the negative impact of K&F trips was discussed in 2006: the research team calculated an extra distance concerning such trips, which was 36% more than in the situations if passengers park their cars; K&F zone was mentioned only briefly as an express pick-up area when discussing the possibility to introduce charges [13]. Budd and co-authors in the article [16] also discussed mainly the problems like congestions and air pollution associated more with K&F journeys than K&F zones itself (they called them drop-off zones). The authors came to the conclusion that from the perspective of increasing environmental challenges, airports should find a way to reduce the share of such trips. Miyoshi and Mason [17] reported that carbon from such modes as ‘drop-off/pick-up’ and ‘minicab’ created greater CO₂ emissions and consequently a bigger cost of environmental damage; K&F zones themselves were not mentioned by the authors at all.

Taking into account the facts that the performance and characteristics of airport ground accesses receive a lot of attention in terms of planning [18], and in turn, K&F zones are an integral part in many airports, surprisingly, the organisation and signage of such zones are not regulated, unified and standardised, which is confirmed by a large amount of varieties met all over the world [15]. On the other hand, such unclearness can provide only additional confusion and time loss, especially for passengers using cars and taxis to reach airports, although they are ready to pay more to reduce the time when reaching an airport ground access [19].

Here, it is needed to add that research works concerning such types of infrastructure elements as a kiss-and-ride zone/area (as written above, K&F zones could be considered its variety) also do not cover issues of road users and confusing situations met by them. It should be pointed out that scientific articles related to the organisation and usage of kiss-and-ride zones focus mainly on aspects associated with their locations [20] and quantitative performance parameters (e.g., the amount and types of cars using them, the time of usage by different types of vehicles, etc.) [21].

2.2. Eye-Tracking Technique and Road Signage

The general motivation for choosing, for assessment and verification, an eye-tracking technique as a method to search and estimate the confusion points and problems of drivers was the following: driving a vehicle by its nature may be attributed to cognitive processes, which is founded mostly on visual input and consequently may be measured by the movements of driver’s eyes [4]; this statement could be extended to other types of road users. For instance, when working in the system ‘driver-vehicle-road-environment’, proper scanning of the road environment is of the utmost importance and, in case of failure, may be considered a root of the prevalence of accidents [22].

As it was declared above, the analysis of the K&F zone was focused on its traffic organisation including road signage—road signs and markings—as its key and integral parts. Therefore, it should be pointed out that road signage as well as its elements should be readable and understandable for drivers; at the same time, it must be kept in mind that the detection and comprehension of road signs depends on such factors as the characteristics of road users, environmental conditions and the characteristics of the road signage itself [23]. There are studies confirming problems and differences among drivers when understanding road signs [24,25]. The same could theoretically be applied to road markings, although appropriate research works were not found.

Although there is quite a large amount of eye-tracking studies concerning road signs’ perception in laboratories [26,27,28,29,30], eye-tracking field tests focused directly on road signs and markings, which try to solve different issues, are not common. First of all, it should be pointed out the number of studies carried out by one research team, which confirmed that the eye-tracking technique returns valuable outputs when analysing road signs’ perception by drivers, which will be discussed below [31,32,33]. A systematic review concerning road marking and its influence on drivers’ behaviour and road safety mentioned only two studies, which used eye-tracking tools (both carried out in the field) [34]. Additionally, there are research works describing the usage of the same technology to assess the perception of horizontal road marking by drivers ahead of intersections, which also return sensible outcomes [35,36]. These studies used similar theoretical eye-tracking background and metrics, but all of them were oriented to the general assessment of certain signs without the estimation of traffic organisation itself, although this could be conducted applying the same standard approach.

Taking into account all the above-mentioned points, it should be repeated that this study is addressed first of all to cover the following research gap: to assess and verify the suitability of eye-tracking tools to search and analyse confusion points, obvious and non-obvious problems and challenges encountered by drivers while performing standard activities in such a complex infrastructure object as the K&F zone as well as their possible reasons. There are also two additional aspects captured by this research work, which could be considered as added values: (1) an assessment on how drivers perceive standardised and unstandardised signage in the K&F zone and (2) a comparison of drivers’ behaviour relating to their familiarity with an investigated area. To the best of our knowledge, there is no research work concerning K&F zones conducted from the perspective of drivers’ perception; also, this is the first of such a study carried out in Poland. Only standard eye-tracking measures were applied within the scope of this study—the reasons for such a decision will be discussed below.

3. Field Study

3.1. Analysed Areas

The considered K&F area is located in the airport surface access of Kraków Airport. Taking into account the fact that to reach this part of the landside area as well as the road network organisation near the airport itself could also create a challenge for drivers, it was decided to consider the approach to the airport ground access separately when planning the field study. The scheme in Figure 1 shows the approach to the airport landside marked as Zone A and the airport ground access marked as Zone B. The boundary between the two zones is associated with the moment when drivers start to see the entrances to the landside of Kraków Airport. The scheme includes also a planned travel path trajectory and, as was mentioned above, important elements of the traffic organisation—road signs dedicated for drivers along this route as well as road signs, which are visible when driving the planned route.

Here, it should be clarified that the whole travel path was around 8.5 km from the starting point: the 7.5 km way to the test field should be considered to drivers as a time to become familiar with the equipment—the eye-tracker and vehicle—and stop paying attention to the first one. The route stretch through Zone B has a length of 0.450 km with 56 road signs (124.4 signs per km) and clearly represents an area related directly to the K&F zone and their organisation. As it was written above, Zone A should be considered as an important element of the route: it is situated just before Zone B, has a length of 0.295 km and is also full of road signs (13 signs a with density of 44.1 signs per km). Such high densities in both zones create a higher cognitive load on drivers and consequently could be a reason for additional confusions.

3.2. Equipment and Participants

The data were collected using Tobii Pro Glasses 2 (Tobii AB, Danderyd, Sweden) and then processed in the Tobii Pro Lab. The device looks like normal glasses and is easy to wear due to its light weight and slim design. This specific model of eye-tracking spectacles allows recording typical parameters such as gaze, fixation, fixation duration and saccade. For the purpose of this study, the duration of one gaze was set as 0.02 s as well as the duration of one fixation—0.08 s (four gazes). Also, it was assumed that one fixation could be interpreted here as one observation. That is why, within the scope of this study, the words “fixation” and “observation” will be used as synonyms. The Tobii Pro Glasses Controller software version 1.86 was used to calibrate the device for each participant.

The test vehicle was an estate car, with a kerb weight of 1170 kg, equipped with a 1591 cm³ petrol engine and manual gearbox transmission. Distances to the eyes from the ground level and from the hip-point were 120–130 cm and 63 cm, respectively, which is within the standard limits [37,38]. Although Tobii Pro Glasses 2 has a camera installed viewing in the direction of driver’s gazes, and the vehicle was additionally equipped with a front dashboard camcorder (30 frames per second, resolution 1920 × 1080 pixels), which was recording videos of the view in the direction of a car movement. Taking into account that drivers can turn their heads when driving, such a device allowed us to obtain more complete information concerning the situation on the road ahead and its surroundings.

When wearing eye-tracking glasses, the task of participants during the experiment was to reach by car a ground access of Kraków Airport from the starting point, find the K&F zone there and stop in this zone for a moment, imitating the drop off/pick up of passengers. After that, the drivers had to leave the K&F zone and the airport surface access in the direction from which they arrived (i.e., to the east of the airport—the first exit in Figure 1). Driving tests were carried out during July and August of the year 2020 when the air transport traffic all over the world was drastically reduced because of the declared COVID-19 pandemic [39,40]. At the same time, these months were associated with the moment when weakening the pandemic restrictions in Poland. On the one hand, such situation allowed us to make tests, and on the other hand, the traffic at the airport was very low (according to our measurements carried out at the end of July 2020, there were only 77 vehicles per hour for 38 available parking places in the K&F zone; one vehicle occupied a parking slot for 3 min and 10 s on average), which allowed us to realize the scenario when interactions with other road users were almost excluded and tests could be focused on elements of infrastructure and traffic organisation. Such a situation also neutralized the influence of such parameters like speed, acceleration and deceleration of a vehicle, the waiting time for K&F parking place, etc., making their measurement unnecessary. That is why the study concentrates on standard eye-tracking metrics (number of fixations, duration of fixations, etc.). During the driving tests, there were only two types of weather conditions: sunny or cloudy (no precipitation). Here, it should be noted that the possible influence of sunny conditions will be discussed only when entering/exiting the K&F zone: at this time, the K&F zone itself was in shadow of the terminal building, which eliminated any impact. Importantly, the ethical guidelines set by Politechnika Krakowska and Palacký University Olomouc concerning privacy protection were followed by data processing.

After the field test, each participant completed a short questionnaire covering basic demographic data, driving experience, familiarity with Kraków Airport and its surroundings as well as subjective assessments of the route they drove. Moreover, 23 drivers (aged 20–24, average 22.4; 16 males and 7 females) were classified into 3 groups based on the number of visits at Kraków Airport:

• Group I—came to the airport for the first time during the study (3 males and 4 females);
• Group II—had been there only once before the study or in the last few years have used the airport more than once (10 males and 3 females);
• Group III—visited the airport often because of their work (3 males).

Not only drivers’ ages but also their driving experiences were quite similar: at the moment of the field study, participants had a driver’s license for an average 4.2 years (range 2–6), and all of them drove regularly.

3.3. Assessment of Road Signs

As it was declared above, the main focus of this study is signage and especially road signs. To better understand the area of the field study, all road signs along the travel path were assessed by using a scale presented in

Table 1. Subjective assessment scale for road signs.

Rating	Correctness	Visibility	Quality of Sign Face
3	Appropriate	Good, no obstructions	Good, clearly legible
2	Counterintuitive, information unclear, unnecessary, redundant	Poorly positioned, possibility of not noticing, crooked or spaced inappropriately	Adequate, not affecting overall visibility or legibility, but road sign requires replacement
1	Erroneous or inappropriate, confusing	Positioned incorrectly, could become hard to notice (partially obscured or without contrasting background), confirmation sign on left side positioned improperly	Damaged, poor quality, faded
0	Missing road sign when required	Obscured or invisible	Missing road sign when required

. It should be mentioned that to make this assessment, recordings from the dashboard camcorder together with field observations were used. This protocol was used before and is considered by authors to be reliable [41,42]. Only one simplification was carried out: signs situated on the same post were assessed together because they were always observed together by participants.

The result of the estimation is presented in

Table 2. Results of road signs assessment.

Zone	A	B
Number of all road signs	13	37
Average rating (all signs)	2.97	2.58
Correctness	3.00	2.35
Visibility	2.92	2.38
Quality	3.00	3.00

. Here, the relatively low rating of signs in Zone B should be noted. It seems to be that the main reason is their correctness and visibility. This indicates possible problems, which drivers could encounter while moving to and in the K&F zone.

4. Overview of Road Signs Perception

An overview of the road sign observations is presented in

Table 3. Road signs observation by driver groups.

Group	All Drivers	Group I	Group II	Group III
Number of observed signs per person: average (range)	10 (3–23)	14 (7–23)	8 (3–23)	8 (5–10)
Number of observations per person: average (range)	17 (4–50)	27 (13–50)	13 (4–33)	10 (7–15)
Route difficulty (a) (standard deviation)	4.1 (0.7)	4.1 (0.9)	4.0 (0.7)	4.0 (0.0)

^(a) Subjective assessment of the route difficulty after the ride on a Likert scale of 1–5, where 1 is demanding and 5 is effortless.

. These results show only one clear difference between drivers who have never been at Kraków Airport and the rest of the drivers: the drivers from Group I observed around 40% more road signs as well as had twice as many observations (fixations). For both parameters—the number of observed signs per person and number of observations per person—there is no statistically significant difference between Group II and Group III: the p-values are equal to 0.972 and 0.536, respectively. On the other hand, both these groups differ significantly from Group I: all p-values are less than 0.05 (although it is necessary to remember that the size of Group III is quite small, which could be considered as a weakness within the scope of this calculation). Indeed, the unfamiliarity with the test route could be the most obvious explanation of such differences.

Interestingly, when assessing the difficulty of the route, all groups of drivers put in a similar grade. This confirms that, although drivers did not feel it was problematic to reach the K&F zone and move through it, their subjective perception did not fully agree with the objective issues and challenges met during the field experiment, which will be discussed in the next chapter. Another explanation could be that local drivers are accustomed to constant problems on Polish roads while driving which are associated with errors, inconsistencies, uncertainties, visibility difficulties and the redundancies of road signs. Of course, such a hypothesis requires additional wide research, which is out of the scope of the presented study.

Before coming to details concerning the observations of road signs, it should be pointed out that no significant differences in fixations and their durations among different groups of participants were found. That is why there was no division into groups of drivers when discussing the observations of signs in this chapter, but this thread will be returned to in the chapter 5 when presenting challenges met by drivers in the K&F zone.

Also, it should be mentioned that in comparison with previous research works associated with road sign observations and perceptions [31,33], it is important to point out a difference in fixation duration: the current study showed much higher values. This could be explained by the difference in test fields (this research team made driving tests on secondary/rural roads with one lane in each direction) and mainly by the high number of road signs per kilometre. Previous studies considered 75 vertical road signs per 8.34 km (9.0 sign per km) with an average fixation duration of 0.154 s and 69 vertical signs per 27 km (2.5 sign per km), with fixation durations from 0.10725 s to 0.16932 s, depending on driving skills. As was mentioned above, the current study consists of two stretches: 0.295 km with 13 signs (zone A) and 0.450 km with 56 signs (zone B). In the approaching road segment with a road sign density of 44.1 signs per km, fixation durations vary from 0.34 to 0.55 s, and in the K&F zone, with 124.4 signs per km, fixation durations were much higher—0.19–1.28 s.

4.1. Approach to Airport Surface Access

The scheme in Figure 2 presents how often all road signs were observed by test drivers in Zone A.

Table 4. Signs most frequently observed in Zone A.

Sign Picture	Sign Number	Observers (Percentage)	Observations (per Observer)	Observation Distance [m] Average (Range)	Observation Time [s] Average (Range)
	1	10 (43%)	20 (2.0)	22 (14–35)	0.34 (0.08–0.96)
	2	6 (26%)	8 (1.3)	40 (20–72)	0.55 (0.12–1.26)
	3	20 (87%)	66 (3.3)	37 (2–125)	0.43 (0.08–1.76)
	4	13 (57%)	16 (1.2)	24 (5–59)	0.41 (0.18–1.48)
All other road signs (six signs)		8 (35%)	18 (2.3)	22 (6–53)	0.25 (0.10–0.58)

shows signs which were observed more frequently than others in Zone A: they are also labelled in Figure 2. Although the observation frequency of signs 2 and 4 seems to be logical, because they helped the driver to choose a movement direction in the lanes, sign numbers 1 and 3 also made them to the list of the most observed ones. The possible reasons for such a situation could be (1) the biggest amount of information, which they present in comparison with other signs in Zone A, and (2) the aim of both signs is to help drivers when finding the way/direction to the airport. Also, it should be noticed that average fixation durations are quite similar for all of the most frequently observed signs in comparison with other ones. This is quite surprising considering the standard form of sign numbers 2 and 4.

4.2. K&F Zone

Figure 3 presents the scheme depicting how often all road signs were observed by drivers in Zone B.

Table 5. Signs most frequently observed in Zone B.

Sign Picture	Sign Number	Observers (Percentage)	Observations (per Observer)	Observation Distance [m] Average (Range)	Observation Time [s] Average (Range)
	5	6 (26%)	8 (1.3)	19 (10–34)	0.19 (0.08–0.40)
	6	6 (26%)	8 (1.3)	60 (26–106)	0.27 (0.08–0.94)
	7	13 (57%)	29 (2.2)	39 (3–113)	0.44 (0.14–1.88)
	8	11 (48%)	19 (1.7)	35 (9–80)	0.54 (0.08–1.88)
	9	8 (35%)	8 (1.0)	67 (18–113)	1.28 (0.18–4.10)
	10	13 (57%)	27 (2.1)	30 (10–67)	0.38 (0.08–1.26)
	11 (a)	13 (57%)	21 (1.6)	23 (5–56)	0.35 (0.08–0.94)
	12	7 (30%)	9 (1.3)	55 (26–89)	0.47 (0.12–1.00)
All other road signs (33 signs)		21 (91%)	110 (5.2)	28 (2–93)	0.38 (0.08–2.56)

^(a) Sign 11 presents regulations for using the K&F zone.

shows details of the road sign observations in Zone B. Firstly, it should be pointed out that there was a large number of fixations on road signs, which were dedicated not for drivers moving along the planned travel path but for drivers who moved along the parallel road leading directly to long-term car parking. This confirms that the number of road signs situated in the comparatively small area of the airport surface access is quite a large (124.4 signs per km), and they are located very close to each other. An additional discussion concerning sign number 12 will be provided below. Secondly, in comparison with Zone A, the distribution of fixation durations is quite spread in Zone B—0.34–0.55 s vs. 0.19–1.28 s. Again, this could be associated with the large amount of road signs located there: drivers had to switch more often between them when searching for the necessary information.

Additional comments are required for the two signs from Table 5: number 5 and 11. Sign number 5 consists of three elements, one of which is ‘Stefa ruchu’ (in English: ‘Traffic zone’): this sign informs drivers about the fact that all general traffic regulations should be applied on the marked road, even if it is an internal road, which, in this case, constitutes the road and its infrastructure of the airport ground access. Sign number 11 (as well as sign numbers 7, 8 and 9) is not a part of the standard signs according to Polish traffic law: such a sign should be developed individually and approved by an appropriate authority responsible for roads where they plan to be installed. According to the outcome of the study, this sign was frequently observed with a relatively long average observation time. This means that drivers spent much time trying to read its content. Unfortunately, sign number 11 was hardly visible for drivers because of its small font. Therefore, it can be concluded that, instead of being helpful, this sign mainly created confusion (some drivers considered it even as an ‘information noise’—an excessive piece of data which arose mainly because of formal requirements).

5. Confusion Points of K&F Zone of Kraków Airport

Here, we discuss the confusion points and troubles met by drivers associated mainly with inadequacies in the signage of the traffic organisation at the K&F zone of Kraków Airport, which were identified and assessed with the help of the eye-tracking technique. In general, problems were noticed in each part of the area: at the entrance and exit as well as in the zone itself. The data extracted when processing videos from the dashboard camcorder were used only as an additional way to verify and confirm the outcomes obtained from the eye-tracking data.

5.1. Entrance to the K&F Zone

When considering the entrance, drivers have two options to reach the K&F zone by choosing one of two gates, which are marked separately by signs above the road: the left one—sign number 8—and the right one—sign number 7 (Figure 4). Both of them were among the most frequently observed ones (Table 5). When crossing the right gate (Figure 5b), the road section leading to the K&F zone begins with another type of surface (paver blocks instead of asphalt), but it seems like it should not be a big problem for drivers to find the way to the K&F zone, although there are no perceptible road signs, which help to reach it. Another situation arises with drivers using the left gate.

Drivers who choose the left entrance gate in accordance with sign number 8 do not get any signals that they should change lanes to the right one to reach the K&F zone—the sign suggests that they should continue to move directly ahead. In addition, horizontal road markings on the lane (arrows) are confusing because they also indicate the direction of straight-ahead movement (Figure 5a).

Besides the above-mentioned road marking arrows, a dashed line dividing lanes from the left and right gates can create a separate confusion. This dashed line was painted on the edge of paver blocks, which, under sunny (glare) conditions, makes it look like a continuous line or even blends into such a surface. Because drivers recognize road markings based on the visibility level, which is provided by the contrast against the neighbouring surface [43] and could be estimated by using the contrast ratio (CR) measured through luminance [44], the calculation of the Michelson CR (CR_M) was carried out to confirm this effect here. As shown in Figure 6, three areas were selected: the painted part of the dashed line as a white marking, the road surface between painted parts as a space and road surface of paver blocks near the dashed line as a background. Figure 6a presents cloudy weather conditions as a good one and Figure 6b—sunny weather with the presence of sunshine glare, which is considered as critical visibility conditions [44,45]. When calculating the CR between a white marking and a space, the CR_M values are 0.05 and 0.00 during good visibility conditions and glare, respectively. Such small values confirm the fact that the considered dashed line could look like a continuous one, especially in the glare conditions. The CR_M values between elements of the dashed line and a background are also very low: 0.10 and 0.06 for good and glare conditions, respectively. This means that horizontal road marking could really blend with the surface (as, for example, can be seen in Figure 6b with the right edge line and arrows—the CR_M values in both cases are 0.00) and in such a way create a problem for drivers.

During the driving tests, eight participants encountered weather conditions with sunshine glare interference: three, four and one from Groups I, II and III, respectively.

Table 6. Average number of fixations on road markings per driver.

Group	All Drivers		Group I		Group II		Group III
Weather Condition	Cloudy	Glare	Cloudy	Glare	Cloudy	Glare	Cloudy	Glare
Left edge line	0.7	1.3	1.3	1.0	0.6	1.5	0.0	2.0
Right edge line	0.2	0.8	0.8	1.7	0.0	0.0	0.0	0.0
Arrows	2.3	1.7	2.5	1.3	2.1	0.5	3.0	5.0

shows the average number of fixations focused on road marking elements for all groups, excluding two drivers from Group II who used the lane from the left entrance gate (their cases will be described separately below). Therefore, Table 6 includes only drivers who entered the right gate and did not switch lanes.

If the ‘lack’ of the right edge line surprised participants for Group I and the arrows visible in good conditions attracted more of their attention, then it seems that drivers from Group II concentrated more on arrows and the left edge line when it was cloudy and sunny, respectively, which depended likely on what was more helpful and clear for them to keep in their lane. Even drivers from Group III could be confused a bit: during good visibility conditions, they paid attention only to arrows and only looked at them a few times probably to confirm their driving direction, while sunshine glare interference forced them (1) to observe the arrows more intensively (highly likely searching them because of their low visibility) and (2) to fix their eyes also onto the left edge line (probably because its ‘new’ continuous shape).

Also, it should be added that the reliability of the presented data in Table 6 is relatively weak because of sample size—any overall conclusions about groups are not possible, but even a simple analysis allows us to identify and/or confirm a confusion point and its probable reasons.

Amongst 23 drivers who took part in the experiment, only 1 participant from Group II chose the left entrance to the zone, which was really confusing for him and consequently should be discussed in detail. There were 28 fixations between the start of vehicle movement from the left gate and the change in the lane to the right one, which could be nominally assigned to a few stages. The first 10 fixations can be associated with the orientation stage: the driver observed the road ahead searching for stimuli, which directed him to the K&F zone. During this stage, he also looked at road markings: in the beginning, he looked at the left edge line of the lane (fixation number 4) and then the arrow and the right edge line (fixation numbers 9 and 10). Here, two moments should be pointed out. Firstly, the driver started to move only after the fixation on the left edge line of the lane, which confirms again the importance of road marking for lane-keeping. Secondly, the arrow and the right dashed line, which looked like a continuous one because of sunshine glare conditions, guided the driver to move only straight ahead. Beside road ahead and road markings, there was no stimulus observed by the participants. The second stage can be called the confusion one and included eight fixations. Apart from road ahead observations, the driver focused on the horizon and surroundings to search again for indicators of the K&F zone. At the end of this stage, he looked again at the confusing road marking arrow and then highly likely took a decision to change lanes because the last fixation of the confusion stage was on a road ahead, where there was no stimulus and, at the same time, was the longest one (0.52 s). The last and final stage consisted of 10 fixations and was dedicated fully to the manoeuvre of lane changing. Most of the observations were concentrated on road surroundings and the right car mirror to avoid a collision with other vehicles moving from the right entrance gate. Here, it is important to point out that the last fixation of this stage was performed on the right edge lane directly before the moment when the car crossed it. This fact could suggest that the driver was still not entirely sure whether the manoeuvre of lane changing did not break any regulations. Summarising, this participant entered the K&F zone only at the last possible moment after previously searching unsuccessfully for signs or any information in all possible directions. Probably, the indirect stimulus, which helped the driver here, was either the lack of parking slots on the lane from the left entrance gate or the observation of the right lane made of paving stones and the cars ahead situated along it. Additionally, it should be mentioned that, beside sunshine glare interference, rainy weather conditions could have the same negative impact on drivers: such conditions did not appear during the field test, but, as an example, a wet surface is shown in Figure 7 where the CR_M values are equal to 0.02 and 0.04, calculated for a white marking in relation to a space and a background, respectively.

As mentioned above, it is important to point out again the lack of any perceptible road signs, which inform one about the presence or the beginning of the K&F zone. The only information near the entrance is sign number 11 (provided in Table 5) and boards located on the sidewalk along the K&F parking area, which are completely imperceptible when crossing the entrance gates. This can also lead to mistakes or confusing situations. During the experiment, it was observed that one of the participants from Group II entered the right gate, which leads directly to the K&F zone, changed lanes to the left one, then returned to the right one and finally entered the K&F zone. Looking at the data concerning fixations, such behaviour could be explained step by step. Before starting the movement, the driver focused logically on the horizon in the direction of the K&F zone. Eight fixations later, when the participant came closer to paver blocks, his attention was caught by the left edge line of the road marking where asphalt and paver stones were connected: the dashed line on asphalt looked normal, but on paver blocks, it seemed again to be a continuous one because of sunshine glare. Probably, the driver was confused that such a road marking could somehow restrict a possibility to manoeuvre further if he did not leave the current lane. It was also confirmed by the fact that the participant looked two times at the “continuous” left edge line directly before starting to change to the left lane. But, only two fixations later, when the manoeuvre was not still even completed, the driver noticed a car on the surface of paver blocks, which slowly moved ahead in the direction of the K&F parking area. The participant immediately looked again at the left edge line: it seemed that he was confused momentarily but then started to return to the right lane, confirming such a decision by fixating on the moving car, other cars in the K&F zone, the lane separator, etc. Besides the above-mentioned confusion reasons, this case confirms again the critical influence of glare conditions on the visibility of road markings and consequently on drivers’ understanding of traffic organisation.

5.2. K&F Zone Itself

In the case of the K&F zone itself, the information signs above the divider separating the K&F zone from the roadway leading directly to long-term car parking should be mentioned (Figure 8). Five signs located every 20 m indicate the distance which passengers should cover to reach the second half of the terminal—the area of departures.

Although the inscription ‘Kiss & Fly’ was probably intended to indicate its presence in the zone itself, they could be misinterpreted due to the arrow located on the right. Only once was one of the five signs observed by one driver. Located parallel to the road on its left side, they are mostly invisible from the driver’s perspective. Their location is therefore not fully understood as drivers looking for a parking space on the right are not able to see them. Maybe the situation would be completely different if they were turned towards drivers and put to the right road side because by leaving these signs on the left lane side, they start to be comprehended only for drivers using the roadway leading directly to long-term car parking. These signs seem also not to be intended for passengers because, firstly, they would require a relatively long path of movement, which is not an optimal choice, and secondly, they are located quite far from the sidewalk and in places not dedicated for pedestrian traffic. While their location parallel to the road could have its justification according to the regulations and price lists, in this case, they do not fulfil their function at all and invisible signs cannot be the basis for any requirements.

5.3. Exit from the K&F Zone

There are two ways to exit from the K&F zone, the first and the second ones, as shown in Figure 1 (vehicle route map) and Figure 3. The problem is that it is not always clear which exit to take when turning left (Figure 9). According to the conversation with airport employees who are familiar with the traffic organisation project, the first exit from the K&F zone was intended for drivers who want to leave the airport area in the eastern direction; according to the task for drivers participating in the tests, this is the same direction from which drivers arrived and should leave the K&F zone. The second exit was built to lead drivers to the village Kryspinów. Additionally, after turning left, the design of the traffic organisation allows drivers to change a lane and consequently exit.

As it is possible to see in Figure 9, there were two mandatory signs (one on the left, obligating drivers to go only straight ahead and dedicated for drivers who use another lane, and one on the right, which obligates drivers to go only left or straight) and two road marking arrows (one located closer, which is marked with the words ‘WYJAZD’ and ‘EXIT’ and obligates drivers to turn left, and one located further, marked with the letter ‘P’ and obligating drivers to go only straight). All these four elements were observed by participants and could impact on drivers’ decisions when selecting an exit. Here, it is important to point out two aspects. Firstly, the right road sign and ‘turn left’ arrow are formally in conflict, and according to Polish traffic law, the first one has priority—the second one should be ignored by road users. Secondly, the sunshine glare conditions influenced significantly on the visibility of road marking arrows: the CR_M of the ‘turn left’ and ‘go straight’ arrows decreased from 0.20 and 0.13 to 0.14 and 0.03, respectively, which confirm again the impact of road surface materials on contrast (asphalt vs. paver blocks).

Before starting the discussion concerning the choice between two options, it should be noted that there were six cases with another moving car in front of the test vehicle when exiting the K&F zone (two from Group I and four from Group II). Such situations were excluded from further analysis which is presented below because, according to the eye-tracking and observation data, these drivers highly likely followed simply a car ahead.

Only one person from Group I took the first exit: this participant looked shortly at the left road sign, but the next long fixation (0.68 s) was already on the ‘turn left’ arrow—probably, it was a decision moment because, immediately after that, the test vehicle turned left. Four drivers from Group I left the K&F zone using the second exit. In two cases of glare weather conditions, the right road sign seemed to have been decisive: drivers observed only this sign and the ‘turn left’ arrow, which could confuse them, and followed Polish traffic law by ignoring the second one, but unfortunately, their final choice was wrong. A similar situation happened with one participant of this group driving in good weather conditions: the fixation on the right road sign was long (0.62 s), but the next fixation on the ‘turn left’ arrow was even longer (1.66 s), which confirms the appearance of confusion for this driver at this moment and then probably led to the selection of the second exit. The second participant from Group I, who drove during good conditions, faced quite another type of confusing situation. This driver observed only horizontal road markings: there were two fixations on the ‘turn left’ arrow and then one fixation on the ‘go straight’ arrow—the driver decided to move according to the last one. Summarising, all described cases confirm that it is impossible for drivers who visit this place for the first time to make a clear decision based on the existing signage.

There were two drivers from Group II who left the K&F zone in sunshine glare conditions. Both of them did not focus on any of four considered elements of the traffic organisation but searched for other stimuli. Unfortunately, perhaps due to these unfavourable weather conditions, nothing attracted their attention, and that is why they missed the first exit and used the second one at the last possible moment. When speaking about good conditions, first of all, we should mention one driver who also did not observe any traffic organisation stimuli and also had a similar course of action with the same result—the choice of the second exit. Three participants of Group II behaved similarly and logically: each of them conducted only one fixation on the ‘turn left’ arrow and then consequently turned left using the first exit. One driver looked only at the ‘go straight’ arrow and went to the second exit, which also could be considered as a logical behaviour. Apart from logical cases, there were also two confusing situations. In the first case, the driver looked first at the ‘turn left’ arrow and then at the ‘go straight’ arrow: the first exit was the final choice, but according to the eye-tracking data, the participant’s doubts about such a decision are visible. In the second case, the driver had only three fixations on the ‘turn left’ arrow, but he finally selected the second exit: looking through the data, it is possible to conclude that in comparison to the other participants, this driver was paying attention to many more details of road surroundings and probably, in such way, missed the first exit. Summarising, even drivers familiarized with the investigated area still met problems and confusions, which confirms the necessity to improve the situation with the traffic organisation and signage.

All three drivers who knew the route and the airport very well (Group III) chose the first exit regardless of sunshine glare conditions: two out of three of them looked at the ‘left arrow’. This confirms that drivers’ experience supports mainly the choice of the first exit, which is intended to be a logical one according to designers. Therefore, such a situation differs from the concept of self-explanatory roads [46].

A separate issue of the exit from the K&F zone is road sign number 12 (Figure 10). Although this sign is dedicated for drivers using the roadway leading directly to long-term car parking, it was observed by seven drivers from both Group I and Group II. Maybe it attracted their attention because of its size and/or its configuration, but in any way, such a situation should not be considered a normal one.

6. Conclusions

The presented eye-tracking study carried out in such a complex infrastructure element as the K&F zone of Kraków Airport allowed us to clearly identify and assess its confusions and problems together with pointing out their possible reasons. Therefore, it confirmed the suitability and usefulness of this technique to provide such a type of work even when applying only its standard metrics. Also, it was confirmed that, firstly, the traffic organisation influences crucially the drivers’ behaviour, and secondly, the correctness and visibility of traffic signs and road markings regardless of weather conditions could help to solve the misunderstanding during the usage of complex infrastructure objects.

As the first added value, the drivers’ perception of standardised and unstandardised signage using the example of the K&F zone was assessed clearly. This assessment led to an obvious conclusion: clear and common regulations of K&F zones including signage and terms of use within at least the whole country can for sure improve the situation and solve major problems and confusions. The second added value of this study to compare the drivers’ behaviour depending on their familiarity with an investigated area was also realized successfully. Additionally, the negative impact of sunshine glare conditions on drivers’ perception of road signage was confirmed again.

Finally, this study discussed the confusion points of Kraków Airport. Although, on the one hand, they could be considered only as local problems which require tailor-made solutions, on the other hand, they show the scale of challenges that drivers could meet in K&F zones. Taking into account the above-mentioned research works discussing different negative impacts associated with such zones, this scale could be projected to all airports which have such an infrastructure element.