We propose a new method of presenting virtual objects of low attention that can affect the walking speed of pedestrians. We call the proposed virtual objects of low attention a Pedestrian Pace-maker Light, hereafter referred to as PPML. The advantages of PPML are that pedestrians can keep clear visibility of the frontal area for safety concerns in situations where PPML is presented in their view through AR glasses. PPML is a set of multiple lighting objects flowing in the direction of travel. Pedestrians perceive a sense of self-motion in the direction of travel through the vection effect caused by the PPML. The experiment was conducted to investigate whether the proposed PPML affects the walking speed of pedestrians or not. We have experimented with the six patterns of PPML. We confirmed that it can affect the walking speed of pedestrians.
Pedestrian assistance [
Speed is an important factor in walking. If an
intelligent and gentle user interface can affect
walking speed, a smooth flow of pedestrians
would be achieved in our society. If it can let the
walking speed down the walking speed, it could
improve the safety of pedestrians. Research on
pedestrian assistance using augmented reality to
affect walking speed and direction has been done
[
In the case of augmented reality intelligent assistance systems, the effect of virtual objects and the clear visibility of the frontal area of the pedestrian should happen simultaneously. So, the presented virtual objects should not interfere with the visibility of the front area for safety concerns in walking. Note that humans cannot gaze at two objects simultaneously [14]. The virtual objects to be placed in the pedestrian view should attract low attention and keep the clear visibility of the frontal area of pedestrians. Yet the low attention should be sufficient to make influence on their walking.
We propose a new method of presenting virtual objects of low attention that can affect the walking speed of pedestrians. We call the proposed virtual objects of low attention a Pedestrian Pace-maker Light, hereafter referred to as PPML. The advantage of PPML is that pedestrians can keep clear visibility of the frontal area for safety concerns in situations where PPML is presented in their view through AR glasses. PPML is a set of multiple-lighting objects flowing in the direction of travel. Pedestrians perceive a sense of self-motion in the direction of travel through the vection effect produced by the PPML. The walking speed of pedestrians can be affected by adapting the speed of PPML.
We compose the PPML with a set of multiple tiny cubic virtual objects. The PPML does not interfere with the visibility of the front area of pedestrians. Pedestrians can easily avoid obstacles and check their surroundings even when the PPML is presented.
The experiment was conducted to investigate whether the proposed PPML method affects walking speed or not. We used two shapes and colors of PPML in the experiment; a vertical, green cuboid and a horizontal, white cuboid. In addition, we set three speeds of PPML in the experiment; slow, medium, and fast. We compared the performance of these six types of PPML and presented the results and discussion.
When using augmented reality to present virtual objects, the visibility of the front area of pedestrians should be kept clear, and the virtual objects should not attract too much attention. These two factors are important for pedestrians to walk safely. When presenting virtual objects, there exists an occlusion problem in which virtual objects overlap with the real world [15].
There are studies in which virtual objects affect
walking speed and direction, but the visibility of
the frontal area for safety concerns was not
discussed [
There is an approach to reduce the attention to virtual objects. In the study of navigation by Tamura et al. [16], they propose a method called "Active Patterns." This method uses multiple objects flowing in the guiding direction. In their experiment, the gaze information of pedestrians was analyzed, and a questionnaire evaluation of the understandability of the navigation was conducted. As a result, it was found that both navigational clarity and walking safety can be achieved.
Our idea of PPML is inspired by the Pace-maker light for cars. The Pace-maker light has been confirmed to affect the driving speed of cars and can potentially improve traffic flow [17,18,19]. In the study by Igaki et al. [17], simulations were conducted in an environment with Pace-maker light placed on the roadway, and it was confirmed that Pace-maker light contributes to reducing the number of traffic jams. In the study by Yanagihara et al. [18], it is confirmed by simulations that cars actively follow the flow when Pace-maker light is placed on the roadway. In the study by Endo et al. [19], it was confirmed that the Pace-maker light reduces the overall traffic jam by about 20% on an actual highway.
We propose PPML as a new method to both effects the walking speed of pedestrians and keep clear visibility of the frontal area of pedestrians. To manage these two factors, a set of multiple tiny cubic virtual objects flowing in the direction of travel is selected as the PPML. Therefore, pedestrians can keep clear visibility of the frontal area as the virtual objects are too small to occlude the real-world objects in their view. In addition, pedestrians can recognize PPML in their peripheral vision and perceive a sense of selfmotion by presenting virtual objects on the side of the frontal area. Figure 1 shows a snapshot of a vertical, green PPML flowing in the direction of travel.
The reasons for this design of the PPML are as follow. The human visual field is divided into two regions; central vision and peripheral vision [20]. The central visual field is used to observe objects and targets directly and is superior in recognizing detailed information. On the other hand, the peripheral visual field is a region that extends outside of the central visual field and is superior in recognizing movements and changes in a wide area. Considering these characteristics, we design the central vision for checking the walking environment of pedestrians. In peripheral vision, the pedestrians can perceive a sense of selfmotion in the direction of travel by the PPML because recognizing moving virtual objects in the peripheral vision maximizes the vection effect [21].
As for the shapes of the virtual objects that compose the PPML, pedestrians may pay more attention to the PPML if its shape is complex. Therefore, we adopt a simple shape for the PPML. A study [16] exists that investigated the effects of three simple shapes, a sphere, a cuboid, and a capsule, by flowing them in the direction of travel. This study confirmed no significant difference in the effects of the three. Therefore, the cuboid is adopted as the simple shape in this study.
Two kinds of colors are used for the objects that compose the PPML. One is green, which has a relaxing and reassuring effect on pedestrians [22]. The other is white, which is neutral for ordinary urban scenes. In this study, we created four types of PPML with different shapes and colors, as shown in Figures 2 to 5.
• Horizontal white PPML (Figure 2) • Vertical green PPML (Figure 3) • Horizontal green PPML (Figure 4) • Vertical green PPML (Figure 5)
We implemented the PPML system to affect the walking speed of pedestrians in the following environment.
Lumin SDK: 0.25.0
Lumin OS: 0.98.35
Magic Leap 1[23] is an optical see-through type head-mounted display that can be used while walking. It has a freshening rate of 120 Hz and a viewing angle of 30° vertically and 40° horizontally. The system is developed using Unity [24], and the application program runs in standalone mode on Magic Leap 1. Figure 6 shows a snapshot of Magic Leap 1.
• • • • • • •
We plan to affect the walking speed of pedestrians by adjusting the speed of the PPML. When we plan to slow down the walking speed of pedestrians, we present slower PPML than walking speed. Similarly, if we plan to increase the walking speed of pedestrians, we present faster PPML than walking speed.
The difference from the normal walking speed of pedestrians is important in the selection of the speed of the PPML. Here, three different speeds of PPML are chosen; slow, medium, and fast speed. Because the participants in the experiment were males in their 20s, the medium speed was set based on their average walking speed. The speed values are set as follows.
4.2.
We initially prepared four types of PPML with different shapes and colors, as shown in Figures 2 to 5. Before the experiment, we conducted a preliminary experiment to select the two types of PPML to avoid a long experiment procedure that may result in unexpected results due to the feeling of getting tired and having an unconscious attitude to the experiment.
In the preliminary experiment, we evaluated the effects of the four PPML types on walking speed by presenting the slow speed PPML. After the experience, we put questionnaires and interviews. The experimental results confirmed that the horizontal, white PPML shown in Figure 2 has the lowest effect of slowing down, and the vertical, green PPML shown in Figure 5 has the highest effect of slowing down. Therefore, we picked up these two for the main experiment.
As for the speed of PPML, we chose the three different speeds described in section 3.3. In addition, the following six PPMLs are selected for the main experiment. Hereafter, the PPMLs are referred to as follows.
• Slow-H-White (SHW): Slow Horizontal and White PPML • Medium-H-White (MHW): speed, Horizontal and White PPML speed,
• Fast-H-White (FHW): Horizontal and White PPML • Slow-V-Green (SVG): Slow Vertical, and Green PPML • Medium-V-Green (MVG): speed, Vertical, and Green PPML • Fast-V-Green (FVG): Vertical, and Green PPML
The main experiment was conducted to investigate whether the proposed PPML method affects walking speed or not.
The participants are prepared with adult men to adopt one average speed of walking throughout the experiment [25]. We experimented with 12 male participants in their 20s who were familiar with the presentation of virtual objects using an optical see-through head-mounted display. All the experimental procedures followed our university rule, and it is certified by the ethics review committee.
We explained the PPML to the participants beforehand and demonstrated the proposed method to familiarize them. Then, each participant experienced the six PPMLs in different presentation orders. Each session includes a short walk for around 12 meters followed by a set of questionnaires shown in Table 1. After all questionnaires were completed, we had a certain time of interviews to obtain the subjective opinions of the experimental participants.
The experiment was conducted in the corridor environment shown in Figure 7. We measured the time of a 10-meter walk along the corridor. The difference between the time taken to pass the 2meter point and the 12-meter point is used to measure the walking time. As a comparison, we also measured the time required to walk 10 meters without the presentation of the PPML. The questionnaire items are shown in Table 1, and the participants rated the three items on a 7-point scale. Note that the actual questionnaire form is written in Japanese, which is the native language of the participants.
We calculated the mean and standard deviation of the obtained data and showed them in Figures 8 to 10. The vertical axis of Figure 8 shows the time required for a 10-meter walk, and the vertical axes of Figures 9 and 10 show the 7-level evaluation. The horizontal axis indicates the score without presenting PPML and the ones of the six PPMLs. A one-sided t-test is performed on the data. In the graph, [*] indicates a significant difference of 5%, and [**] indicates a significant difference of 1%.
Figure 8 shows that there is a significant difference of 1% between the fast and medium speeds at the same vertical green PPML. The same happens between the fast and medium speeds at the same horizontal white PPML. In addition, there is a significant difference of 5% between the fast and slow speeds at the same vertical green PPML, and the same between the fast and slow speeds at the same horizontal white PPML. It is confirmed that PPML affects the walking speed of pedestrians. On the other hand, there is no significant difference in the shape and color of PPML.
Figure 9 shows a significant difference of 1% for all the cases of PPML speed. It is confirmed that the slower the PPML speed is, the slower the pedestrian’s PPML perceived speed is. As for the shape and color of PPML, a significant difference of 5% is found only between horizontal white PPML and vertical green PPML at the same slow speed. It is confirmed that the perceived speed of PPML in horizontal white PPML may be slower than vertical green PPML.
Figure 10 shows a significant difference of 5% between horizontal white PPML and vertical green PPML in terms of the visibility of PPML. It is confirmed that the visibility of vertical green PPML is higher than horizontal white PPML.
Figure 11 shows that there is no significant difference in evaluating the difficulty in walking by presenting PPML and that they are equally evaluated. In addition, it was confirmed that the PPML gave pedestrians no difficulty in walking.
From Figure 8, it is confirmed that PPML affects the walking speed of pedestrians. In addition, it is confirmed that there was no significant difference in the effect on the walking speed between horizontal white PPML and vertical green PPML. On the other hand, in the interviews, some participants said that vertical green PPML had a more substantial effect on the walking speed of pedestrians than horizontal white PPML. Therefore, further experimentation would be needed to investigate how the PPML can affect walking speed regardless of its shape and color.
From Figure 9, it is confirmed that the slower the PPML speed is, the slower the pedestrian’s PPML perceived speed is. In addition, it is confirmed that the perceived speed of PPML in horizontal white PPML may be slower than vertical green PPML. In the interviews, some participants said that some participants perceived vertical green PPML to be faster than horizontal white PPML. Therefore, there were no discrepancies between the interviews and the experimental results.
From Figure 10, it is confirmed that the visibility of vertical green PPML is higher than horizontal white PPML. In the interviews, some participants said that horizontal white PPML is more challenging to see and less present than vertical green PPML. Therefore, there were no discrepancies between the interviews and the experimental results.
From Figure 11, it is confirmed that participants evaluated the perceived walking difficulty equally when presented with vertical green PPML and horizontal white PPML. In the interviews, some of the participants said that they did not feel any difficulty walking with both PPMLs. Therefore, it is indicated that the proposed method may be less likely to give pedestrians a sense of difficulty in walking.
We proposed the PPML as a new method of presenting virtual objects of low attention. This method can both affect the walking speed of pedestrians and keep the clear visibility of the frontal area of pedestrians.
The experiment was conducted to investigate whether the proposed PPML method affects walking speed or not. From the experimental results, it is confirmed that PPML affects the walking speed of pedestrians. In addition, there is no significant difference in the effect on the walking speed of pedestrians between horizontal white PPML and vertical green PPML. It is also confirmed that the proposed method is less likely to cause difficulty in walking.
We think more investigation of the effect of walking speed change by PPML should be conducted. The test of clear visibility for safety concerns should be done to very our proposed approach is effective in practical situations too.
Part of this research is supported by Kaken 21H03476.
virtual humans in AR, in: IEEE transactions postural control, in: Handbook of sensory on visualization and computer graphics, physiology, 8(1978), pp. 755-804. 24(4), 2018, pp.1525-1534. [22] N. Kaya, and H. H. Epps, Relationship doi:10.1109/TVCG.2018.2794074. between color and emotion: A study of [13] A. L. Guinet, G. Bouyer, S. Otmane, and E. college students, in: College student journal, Desailly, Towards an AR game for walking 38(3), 2004, pp. 396-405. rehabilitation: preliminary study of the [23] Magic Leap, Inc, Magic Leap 1, viewed 20 impact of augmented feedback modalities on June 2023. URL: walking speed, in: 2020 IEEE International https://www.magicleap.com/ja-jp/magicSymposium on Mixed and Augmented leap-1.
Reality Adjunct (ISMAR-Adjunct), 2020, pp. [24] Unity Software, Inc, Unity, viewed 20 June 264-268. 2023. URL: https://unit.com/. doi:10.1109/ISMAR- [25] M. M. Samson, A. Crowe, P. L. De Vreede, Adjunct51615.2020.00075. J. A. Dessens, S. A. Duursma, and H. J. [14] J. Palmer, Attentional limits on the Verhaar, Differences in gait parameters at a perception and memory of visual information, preferred walking speed in healthy subjects in: Journal of Experimental Psychology. due to age, height and body weight, in: Aging Human perception and performance, 16(2), clinical and experimental research, 13(2001), 1990, pp. 332-350. pp. 16-21.
doi:0.1037//0096-1523.16.2.332. doi:10.1007/BF03351489. [15] M. M. Wloka, and B. G. Anderson,
Resolving occlusion in augmented reality, in: Proceedings of the 1995 symposium on Interactive 3D graphics, 1995, pp. 5-12.
doi:10.1145/199404.199405. [16] Y. Tamura, H. Shishido, and Y. Kameda,
Evaluation of active patterns on direction instruction for pedestrians, 2022. URL: https://ceur-ws.org/Vol-3297/short4.pdf. [17] T. Igaki, and T. Uchida, Micro traffic simulation to examine congestion-control measures using pace maker lights, Memoirs of the Faculty of Engineering Osaka City University, 59(2018), pp. 7-17.
doi:10.24544/ocu.20190823-003. [18] M. Yanagihara, K. Hiraki, and H. Oneyama,
An analysis of effects on traffic flow considering the difference of following behaviors on moving light guide system, in: JSTE Journal of Traffic Engineering, 6(2), 2020, pp. A_55-A_62.
doi:10.14954/jste.6.2_A_55. [19] M. Endo, H. Nakagawa, M. Fukase, and D.
Hashimoto, The measures against traffic congestion in Tokyo Wan Aqua-Line EXPWY, in: Proceedings of the 34th Traffic Engineering Research Paper, 2014, pp. 255261. [20] H. Strasburger, I. Rentschler, and M. Jüttner,
Peripheral vision and pattern recognition: A review, in: Journal of vision, 11(13), 2011.
doi:10.1167/11.5.13. [21] J. Dichgans, Visual-vestibular interactions:
Effects on self-motion perception and