The Influence of Light in the ArenA Boulevard: Dynamic Visibility to Improve the Quality of Residence

Research paper and Design document for the Interactive Lightstreet on the ArenA Boulevard in Amsterdam. Commissioned by Co-ReUs, in collaboration with the University of Applied Sciences of Amsterdam, in association with Mathijs Haakman and Mitchell Ebbers (Computer Engineering).

Interactive Lightstreet Promo Video (Vimeo)

Abstract

Pedestrians experience a poor quality of residence at the ArenA Boulevard in Amsterdam. Based on a long observation from the roof of the Deutsche Bank building, it can be stated that the pedestrians walk far apart from each other which causes them to feel vulnerable. By implementing an interactive light street in the pavement of the boulevard the walking trails of the pedestrians can be influenced, causing them to feel safer due to a sense of cohesion. Because the pedestrians react positively on applying visual borders it is possible to attract them into walking more closely together. By using several LED-strips and ultrasonic sensors that measure distance and send this to a server using the wireless communication protocol XBee, the LEDs will light up at the places where pedestrians walk. Street lighting has a positive influence on the pedestrians’ perception of safety. However, because the traditional street lantern does not necessarily increase overall sight, but rather increases the contrast between visible parts and not-visible parts, it is important to work towards a dynamic field of view. This dynamic field of view will light up where pedestrians walk, resulting in giving them more insights in the walking trails of other pedestrians. Therefore, it can be argued that by implementing a light street on the ArenA Boulevard that offers a dynamic field of view, the quality of residence of the pedestrians will increase. To test this hypothesis, a modular light street was designed, developed and applied for a series of experiments. From the results of these experiments it can be stated that not only are the pedestrians motivated to use the light street’s dynamic field of view, but also that the light street actually influences the pedestrians’ walking line by an average of 1,47 meters off the original walking line, meaning that the walking lines can be influenced.


Background

The ArenA Boulevard is a part of the Amstel III / Bullewijk-area in Amsterdam. The district is despite its 500 social housing- and 170 student-housing-accommodations mainly an industrial area. The district annually draws about 16 million visitors, whom mainly shop at the IKEA or go out at places like the Ziggo Dome, the AFAS Live venue or the Amsterdam Arena. The total amount of people working in the Amsterdamse Poort or the office-buildings is almost 50.000 (Gemeente Amsterdam, 2016, Gebiedsanalyse Stadsdeel Zuid-Oost).

Related Work

Based on four two-and-a-half-hour long observations from the roof of the Deutsche Bank, time-lapses were collected; from this information walking lines were established. The conclusions drawn out of the walking lines are as follows:

1. The pedestrians use the ArenA Boulevard merely as a passageway and not as a point of destination. Contrary to other squares and public spaces in Amsterdam. The Leidseplein, for example, is mainly used as a point of destination.

2. The pedestrians walk far apart from each other, regardless of direction and point of destination. The East-to-West flow, from the Bijlmer-ArenA Station to the Villa ArenA, is almost equal to the West-to-East flow.

3. The walking lines of the pedestrians can be influenced by applying visual barriers. There is an elongated gutter in the pavement that has been unconsciously seen as a border. The pedestrians that walk along this gutter will not pass the line, even though their point of destination would be reached sooner if they would.

Secondly, an investigation in association with the researchers from Co-ReUs has led to heatmap-visualizations that show the quality of residence-levels according to 165 respondents. These maps show what places in the boulevard are seen as attractive places and which places are not. The visualizations show that the area nearby the station and the shopping-area is experienced as attractive, while the area more west, towards the Villa ArenA, are experienced as not-attractive, unpleasant and unsafe (Co-ReUs, 2017, Co-Creatiesessie 4).

Figure 1: Heat-maps of area-connotation

Earlier research on the psychology and perception of pedestrians in public spaces as shown the following two interesting aspects: (Nasar, Bokharaei, 2016, Journal of Environmental Psychology).

1. As soon as a pedestrian experiences a negative quality of residence in a public space, the person is less quickly inclined to walk next to other, unknown, pedestrians.

2. When a pedestrian walks alone, and not next to others, the person will be likely to experience a negative connotation to the public space.

Based on this research it can be concluded that there is a downwards spiral concerning the quality of residence and the sense of safety. It is possible that this is one of the reasons the Bijlmer-ArenA district carries a relatively high grade when it comes to the Amsterdam insecurity index. The Bijlmer-ArenA scores 122 on the index, which is higher than the Amsterdam average of 108. This index is based on the sense of safety and a higher score means the area is seen as more unsafe by more people (Gemeente Amsterdam, 2016, Gebiedsanalyse Stadsdeel Zuid-Oost).

Light plays an important role when it comes to the pedestrian’s perception of safety, not only has it been proven that street lighting in public space decrease criminal activities (Welsh, Farrington, 2008, Intelligent Street Lighting and Perceptions of Personal Safety), but also that by applying sufficient street lighting the collective sense of safety and security can be improved. Once this sense of safety and security due to sufficient street lighting has been reached, more pedestrians can be found after dark in public spaces (Painter 1994, Painter & Farrington 1999, The Influence of Street Lighting Improvements on Crime, Fear and Pedestrian Street Use, after Dark). Therefore it can be concluded that as soon as the sun sets, street lighting plays a major role in the quality of public spaces (Durak, Olguntürk, Yener, Güvenç, & Gürçınar, 2007; Küller, Ballal, Laike, Mikellides, & Tonello, 2006; Miwa & Hanyu, 2006; Tiller, 1990, Impact of Lighting Arrangements and Illuminances on Different Impressions).

First, It is important to address that the relation between street lighting and the perception of safety still can be seen as quite ambiguous. Although, lighting offers the pedestrians a visible field of sight, light still cast shadows and therefore increases the contrast between light and dark, resulting the contrast between highlighted and shaded areas. This contrast is why a light street that lights up where pedestrians walk can improve the sense of safety, more than the traditional street lantern which is placed all over the boulevard.

The second major difference between the light street and the current traditional street lanterns is that the light street lights up the pavement upwards, contrary to the lanterns that cast light downwards from a tall pole. Earlier research to the quality of residence in public spaces has shown that there is an actual difference between lighting form an overhead position and from a peripheral position. Lighting from an overhead position would cast a smaller radius and a higher contrast and therefore increase sight in the highlighted area. A peripheral positioned light however, would cast a bigger radius with a contrast that is lower than that of an overhead lighting. This same research has shown that a non-uniform peripheral light improves the quality of residence significantly in in comparison with a uniform overhead positioned street light (Flynn, 1988, Lighting-design Decisions as Interventions in Human Visual Space).

Methodology

To execute a series of different experiments, a modular light street must be designed that can be easily adjusted to any context or scenario. The components of the light street must communicate wirelessly and must be able to adapt to sudden changes. For an illustration of the anatomy of the light street, see figure 2.

Figure 2: The anatomy of the light street

(1) Arduino Sensor, (2) XBee Sensor, (3) Battery Sensor, (4) Ultrasonic Sensor, (5) Arduino LED strip, (6) Powerbank LED strip, (7) XBee LED strip, (8) Sodaq, (9) XBee Sodaq, (10) Battery Sodaq, (11) Sensor Case, (12) Casing LED strip (top), (13) LEDs, (14) Casing LED strip (bottom), (15) Cable LED strip to PCB.

The light street consists of four LED strips that are each two meters long with 60 individually controllable LEDs per meter. In the setup for the Arena Boulevard two LED rails were placed. Both LED rails concists of a combination of two LED strips. The surface of the light street was four meters long by two meters wide. Each LED rail is equipped with a battery and a printable circuit board holding a XBee and an Arduino. Because every LED rail is powered and can communicate wirelessly, it can be modularly placed. Also, there is a total of four sensorcases. Every single sensorcase communicates wirelessly with the Sodaq and because each case has been equipped with its own battery they can also be placed modularly.

The ultrasonic sensors measure distances constantly through echo-location. The range of the ultrasonic sensors is between a minimum of 20 centimeters and a maximum of 701 centimeter. When there is no object in the sight of the sensor it will return the maximum-distance of 701 centimeters. These distances will be transmitted to the Sodaq through the XBee module at a frequency of ten times per second. The Sodaq, connected with the XBee, acknowledges either a 1-value or a 0-value to the signal. Whether a 1- or 0-value is assigned is dependant on the set range of the sensorcases. In the setup for the experiments on the ArenA Boulevard the following applied:

Number of LEDstrips: 4
Number of LEDrails: 2
Number of LEDs per LEDrail: 240
Surface light street: 8 m2
Number of sensorcases: 4
Range sensors: 20–220

The sensorcases are positioned with a distance of 20 centimeter from the light street and constantly measure if the signal is 0 or 1. This signal is transmitted to the Sodaq with a frequency of ten times per second.

// the signal is 0 when
sensorValue < 20 || sensorValue > 220)
// the signal is 1 when
sensorValue >= 20 && sensorValue =< 220)

The Sodaq then stores these values in an array. The length of the array is equal to the amount of sensorcases connected. The Sodaq sends the array through the XBee to the LED-rail-Xbee that is attached on the printable circuit boards and turn the corresponding LEDs on. The number of LEDs that go on at a 1-value can be calculated as follows:

XledsActive = (Σ ledStrips * 120) / Σ sensorTotal

In figure 3 an illustration of the top view of the light street can be seen. In this illustration P1 and P2 are shown. These are examples of scenarios where the position of pedestrians, relatively to the light street, is shown.

Figure 3: Top view of the light street

In the case of the scenario of pedestrian P1, the values are:

Sensorcase #1 measures sensorvalue = 442;
Sensorcase #2 measures sensorValue = 701;
Sensorcase #3 measures sensorValue = 701;
Sensorcase #4 measures sensorValue = 701;

Because the person is positioned outside the set maximum range of 220 centimeter, no lights will turn on. In the case of the scenario of pedestrian P2 however, the individual is positioned in the line of sight of sensorcase #3 and therefore the following signal will be transmitted:

Sensorcase #1 measures sensorvalue = 701;
Sensorcase #2 measures sensorValue = 701;
Sensorcase #3 measures sensorValue = 124;
Sensorcase #4 measures sensorValue = 701;
sensorValueOne > 220 // value is 0
sensorValueTwo > 220 // value is 0
sensorValueThree >= 20 && sensorValueThree <= 220 //value is 1
sensorValueFour > 220 // value is 0

The Sodaq is connected with four sensors and an array is created: signal{0, 0, 1, 0}. This is sent to the LEDrails, which will calculate:

XledsActive = (Σ ledStrips * 120) / Σ sensorTotal
XledsActive = 240 / 4 = 60

Hence the array{0, 0, 1, 0}, the result will be as follows: LED(121) to LED(181) will turn on.

Interaction

The light street has two different modes: an Idle-mode and an interaction mode. The idle-mode is activated as soon as there has been no presence in the light street for over 1.7 seconds, in other words an array with {0, 0, 0, 0}. The idle-mode plays a randomizedDots LED-animation, where each time 30 random LEDs per LEDrail will turn on and take a random color.

Each activated LED will be turned on for 2 seconds. The idle-mode is meant to attract pedestrians to the light street from other places on the boulevard. As soon as one of the sensorcases measure a value between 20 centimeters and 220 centimeters, and therefore the array-values change, the Idle-mode will jump to the interaction mode. The interaction mode simply visualizes the position of the pedestrian in the light street, because the LEDs around this person will turn on and follow the person when he or she moves. The interaction, and hereby also the translation of user-input to system-output, plays a major role in the light street.

Because of the interaction, Albert Borgmann’s Device Paradigm theory is very much applicable here. Borgmann states that there has been a huge shift in the way humans interact with their technologies (Borgmann, 1984, The Device Paradigm). This means that the results of using technological applications move further away from the motives to actually use these applications. Thus, the black-box, or the translation between user-input and technology-output, become more complex and less transparant. Borgmann states that this transparency is quite important for the user’s sense of satisfaction. When the black box is fully transparant, the user will understand the relation between user-input and technology-output clearly.

When it can be said that the highlighted areas on the ArenA Boulevard can be manipulated by the pedestrians that walk in the line of sight of the sensorcases, then there is a dynamically illuminated area. However, once the user interacts with the illuminated area, it is important to offer a transparant black-box. The goal is to provide the user with the sense of satisfaction that can be caused by interacting with technology, according to Borgmann’s theory.

Experiments

At the ArenA Boulevard, a bench is placed by the local government of Amsterdam. This bench is positioned in the middle of the boulevard, giving the pedestrians the option to either walk left or right of the bench, creating a binary setup. Whether the pedestrians chooses to walk right or left would not have any impact on the distance walked from the start-point and the point of destination, both options are equally long. To get meaningful results, three experiments were conducted.

The first one is the baseline measurement. Gathered form the observations from the Deutsche Bank, the original walking patterns of the pedestrians in this area were already established. Based on these popular walking patterns a specific walking area was selected and observed for two hours. One week later the experiment was conducted. In this experiment the light street was laid down to the right of the afore mentioned bench, without the sensors and it displayed a constant idle mode. By counting the people who walked via the light street as opposed to walking at the other side of the bench for two hours, a calculation could be made of the difference in pedestrians choosing to walk left or right now the light street is implemented.

In the second experiment, the relation between the user input and the system output was measured. The two extreme conditions in this test are defined as followed:

  1. When a person steps into the light street, the interaction mode changes and illuminates the area around the pedestrian. The person is surprised, but can grasp the logic of the light street. When that pedestrian continues his walk, he sees the lights move with him. He walks till the end of the light street and resumes his normal walking pattern.
  2. When a person steps into the light street, the interaction mode changes and lights the area around the pedestrian. He doesn’t understand what has happened. Just now a different animation played. He stands still, looks around him and takes a step back. Confused he resumes his normal walking pattern.

To get a view of the reason why a pedestrian is confused about the interaction with the light street, there was a small interview conducted with him or her. They were asked about their expectations when they walked onto the light street and why they didn’t anticipate the system output.

The third experiment was conducted to measure the average deviation to the original walking line in meters. First, a popular walking line was measured and assigned a starting point (Pos1) and an end-point (Pos2). The distance between these two points was 60 meters. The light street was positioned slightly outside sky width-line between Pos1 and Pos2. For a visualization of the setup, see figure 4.

figure 4: Visualisation of the experiment-setup

The locations of Pos1 and Pos2 were based on one of the most popular walking lines gathered from the observations. If a pedestrian would walk directly from Pos1 to Pos2, the person would set a distance of 60 meters. However, if the person would walk from Pos1, through the light street and then to Pos2, the pedestrian would set a total distance of 62,34 meters, because the light street is positioned with an eight-degree angle and an offset of 8 meters from the sky-width line between Pos1 and Pos2.

The plan was to conduct the experiments between the beginning of April until the end of June. These experiments ran in the off-peak moments at dusk from 10pm until 12pm. The specific days on which the experiments were held were weekdays and because of the mild temperature of 20°C it can be said that the weather-conditions did not have a negative effect on the amount of pedestrians walking on the ArenA Boulevard on that day.

The expectations from the experiment were that the pedestrians would walk together over the light street near the bench, as opposed to the outer side of the boulevard. Although the light street is quite short for such a large boulevard, the belief is that the light street is a good addition to the ArenA Boulevard as it increases the quality of residence. The pedestrians walking near the light street will spot the object through its illuminated Idle-mode; notifying the pedestrians about the light street.

Results

First of all, a baseline measurement was conducted during the off-peak-hours of the boulevard by doing observations for two hours long. At this point, the light street was not yet placed. From the observations it can be concluded that six pedestrians chose to walk left from the bench and eleven people chose to walk right from the bench. One week later at the exact same time, the light street was placed at the right side of the bench and the same observations were conducted. This time three pedestrians chose to walk on the left side of the bench and twelve pedestrians walked on the right side of the bench and through the light street. By applying an ANOVA-calculation based on these values, resulting in a P-value equal to 0,0474. By choosing the widely used standard of 0.05 for the P-value, we can conclude there is a statistical significance; by placing the light street in a binary setting, it is possible to manipulate the pedestrians’ walking line. This also proves that the pedestrians of the ArenA Boulevard prefer the non-uniform peripheral lighting, compared to the uniform overhead lighting, also known as the traditional street lanterns, at the left side of the bench. However, the limitation of this research is that the time-frame of two-hour observations might have been quite short. The results would have been more trustworthy if the observations would have been longer and more pedestrians would have been observed.

Based on the experiment where the user-interaction was tested, the following insights were gathered. In the darkest, quietest and most desolate area of the ArenA Boulevard a bench is placed by the local government of Amsterdam. This bench stands in the middle of the boulevard, giving the pedestrians the option to either walk left or right of the bench, creating a binary setup. Whether the pedestrians chooses to walk right or left would not have any impact on the distance walked from the start-point and the point of destination, both options are equally long. The results gathered from the first experiment indicate how the walking lines of the pedestrians, and thus the choice to walk either left or right, can be manipulated by implementing the light street.

For experiment three, a four-hour-long observation was conducted and the following insights were gathered: A total of 73 pedestrians walked from Pos1 to Pos2, out of these 73 pedestrians a total number of 46 persons walked through the light street and the other 27 ignored the light street and walked straight ahead.

This means that 46 individuals walked a distance of 62,34 meters and 27 individuals walked a distance of 60 meters. The average distance walked out of these 73 pedestrians is 61,47 meters, which means that the average deviation to the original walking line is 1,47 meters. With this information, it can also be stated that a majority of 64,38% is inclined to undergo a manipulation to its walking line when the light street is implemented in this setting.

Out of the 46 pedestrians that crossed the light street, only five individuals were interviewed about their experience with the light street. Out of these five respondents, three persons indicated that they are convinced of the extent to which the light street improves the quality of residence on the ArenA Boulevard. The other two respondents experienced the interaction with the light street, but have not paid enough attention to the light street to form a proper opinion. As soon as they were addressed on how the light street works, these two respondents were surprised to see the sensor cases. They were not able to form an opinion on how the light street affects the quality of residence.

Future Work

An open and wide space such as the ArenA Boulevard has the advantage of providing the pedestrians with a clear view of the public space, but in this case the disadvantage would be that the light street looks minuscule compared to its surroundings. The result of this is that pedestrians only see the light street when they are very close by. Also, in the distance it is quite hard to recognise the object as a light street. As soon as the light street would be elongated it is crucial to replace the power-source for a stronger equivalent. The current power-source is fit to supply the current installation with power, but is already working towards its maximum capacity.

Because the light street is an outdoor-installation it is important that it can withstand different weather-conditions such as rain, heat, cold, etc. For this prototype the sensor-cases were 3D-printed and are not water-proof. For further research it is suggested to look for a material that is waterproof. Using a more reliable printer will improve the sensor-cases and maybe even make it waterproof when the proper material is being used with professional help. At last, to make the LED-strips more robust and sturdy, it is suggested to cast the LED-strips in a hard-plastic material. The current circumstances cannot withstand lots of pedestrians stepping on the installation for it will break.

Discussion

The experiments have shown that the walking lines of the pedestrians on the ArenA Boulevard can be influenced tremendously by implementing a light street. Even though the first experiment may have had too little respondents to test with. As long as there is a non-uniform peripheral light that provides the user with a dynamic field of sight, the pedestrians are inclined to choose this option over the option of walking through a path illuminated by traditional street lanterns. Therefore, it can be stated that by implementing a light street on the ArenA Boulevard that offers a dynamic field of sight, the walking lines can be manipulated.

However, because the quality of residence is such an abstract and emotional phenomenon it is hard to analyze, measure and address. In this research the factors that influence the quality of residence were collected. These factors include: the sense of safety, the influence of peripheral lighting and a dynamically illuminated area, Albert Borgmann’s Device Paradigm for a sense of satisfaction and the pedestrians’ sense of vulnerability due to walking very far apart from each other. With the experiments conducted, it can only be shown that the walking lines can be manipulated tremendously. The experiments have also shown that the pedestrians of the ArenA Boulevard react positively to the peripheral lighting that provides a dynamically illuminated field of sight, contrary to the traditional overhead lighting.

According to the research on environmental psychology by Nasar and Bokharaei, it can be said that by manipulating the pedestrians’ walking lines in such a way that these individuals would be inclined to walk more closely together, the sense of safety will be improved. This sense of safety will directly influence the Amsterdam insecurity index, causing the index-value to decrease. Therefore, more pedestrians will be motivated to visit the ArenA Boulevard after dark, resulting in a more vivid public space.

However, unfortunately there is still little known when it comes to research about environmental psychology and the quality of residence. To properly validate the increase in quality of residence future experiments must be executed where pedestrians are interviewed about their opinion on the public space. In these interviews the respondents should be asked about their experience with the light street. Also, it would be in favour of the results of the experiments to extend the light street to a total length of 16 meters. This adjustment would require a few technical modifications, such as the improve the power-source so it can supply the elongated light street.

Conclusion

Although it can be stated that the factors that influence the quality of residence can be controlled and manipulated. With the experiment results it cannot be stated yet that the actual quality of residence can be improved.

Even though the walking lines can be manipulated to persuade pedestrians into walking more closely together to decrease the sense of vulnerability, the hypothesis stating that by implementing a light street on the ArenA Boulevard that offers a dynamic field of view the quality of residence of the pedestrians will increase, cannot be confirmed yet. What can be stated though, is that the light street offers a great tool for altering walking lines and can be used for future experiments and research.