저작자표시 - 비영리 2.0 대한민국 이용자는아래의조건을따르는경우에한하여자유롭게 이저작물을복제, 배포, 전송, 전시, 공연및방송할수있습니다. 이차적저작물을작성할수있습니다. 다음과같은조건을따라야합니다 : 저작자표시. 귀하는원저작자를표시하여야합니다. 비영리. 귀하는이저작물

저작자표시 - 비영리 2.0 대한민국 이용자는아래의조건을따르는경우에한하여자유롭게 이저작물을복제, 배포, 전송, 전시, 공연및방송할수있습니다. 이차적저작물을작성할수있습니다. 다음과같은조건을따라야합니다 : 저작자표시. 귀하는원저작자를표시하여야합니다. 비영리. 귀하는이저작물을영리목적으로이용할수없습니다. 귀하는, 이저작물의재이용이나배포의경우, 이저작물에적용된이용허락조건을명확하게나타내어야합니다. 저작권자로부터별도의허가를받으면이러한조건들은적용되지않습니다. 저작권법에따른이용자의권리는위의내용에의하여영향을받지않습니다. 이것은이용허락규약 (Legal Code) 을이해하기쉽게요약한것입니다. Disclaimer

Master's Thesis Hug2Go: The Development of Indoor Smart Driving Personal Mobility Sung-ho Lee Department of Creative Design Engineering Graduate School of Creative Design Engineering, UNIST 2019

Hug2Go: The Development of Indoor Smart Driving Personal Mobility Sung-ho Lee Department of Creative Design Engineering Graduate School of Creative Design Engineering, UNIST

Hug2Go: The Development of Indoor Smart Driving Personal Mobility A thesis submitted to the Graduate School of Creative Design Engineering, UNIST in partial fulfillment of the requirements for the degree of Professional Master of Design-Engineering Sung-ho Lee 07/01/2019 Approved by Advisor Hui-sung Lee

Hug2Go: The Development of Indoor Smart Driving Personal Mobility Sung-ho Lee This certifies that thesis of Hui-sung Lee is approved. 07/01/2019 signature Advisor: Hui-sung Lee signature Committee Member: Cha-jung Kim signature Committee Member: Kwan-myung Kim

Executive Summary This paper presents an initial study on the acceptance of indoor PMVs through providing design and development of a new PMVs (Personal Mobility Vehicles). Hug2Go is the indoor personal mobility, finding passenger through self-driving and going to place by a new way of steering. The personal mobility vehicles (PMVs) emerged as a new category of transportation device in the early 2000s. PMVs offers several potential benefits to consumers and society. Many researchers focused on performance or acceptability of use. However, most of PMVs regarded as outdoor mobility. Recently, popular PMVs has been moved to sharing service area. We thought it opportunity area for the mobility market. In this research, we suppose a new model of indoor mobility and examine it possible to build on the market through the usability test. First, we discovered the context of indoor mobility with existing PMVs driving. Through the observation, we found meaningful insights. Second, we designed and developed indoor PMVs. Third, we conducted a usability evaluation with fifteen participants by using Hug2Go. Experimental results with fifteen participants regarding the acceptance of indoor PMVs validated the proposed latent needs. Finally, we discussed findings and opportunities for improvements. The purpose of this study and the development of PMVs is to provide a comprehensive background for initial research of indoor PMVs. Keywords: PMVs, The indoor mobility, Self-driving, Autonomous vehicle

Table of Contents 1 INTRODUCTION 1.1. Background 1.2. Research Aim and Scope 1.2.1. Research Aim 1.2.3. Research Scope 1.3. Thesis Structure 2 PRELIMINARY STUDY 2.1. Related Works 2.2. Observation to insights 2.2.1. Experimental Task 2.2.2. Pathway 2.2.3. Insights and findings 3 Hug2Go 3.1. Defining target area 3.2. Design features 3.2.1. Three types of operating mode 3.2.2. Switching modes 3.2.3. Sitting on chair (chair mode) 3.2.4. Sitting on a chair facing in the opposite direction (control mode) 3.2.5. The hug steering 3.3. Implementation 3.3.1. Hardware 3.3.2. Motor 3.3.3. Steering 3.3.4. Seat and seat back 3.3.5. Embedded system 4 USABILITY EVALUATION 4.1. Aim & Plan 4.2. Method 4.2.1. Participants 4.2.2. Experimental procedure 4.2.3. Experimental design 4.3. Results 4.3.1. Usage Intention of Hug2Go 4.3.2. Acceptance of the indoor PMVs 4.3.3. Usability of manual driving

4.3.4. Interview 4.4. Findings 4.4.1. Valid PMVs for the indoor space 4.4.2. Incomplete manual steering structure 4.4.3. Not easy but pleasant 4.4.4. Correlation component with acceptance through Factor analysis 4.4.5. Insights through interview 4.4.6. Recognition of moving reverse 4.4.7. A lack of the steering feedback 4.4.8. Emergency button and pedestrian verification 4.4.9. Uncertain affordance (Hug vs. Hold) 4.4.10. Adjusting seat and seat back height or posture 4.4.11. Enhancing convenience 5 DISCUSSION 5.1. Discussion 5.1.1. Validation of the indoor PMVs 5.1.2. Hug vs. Hold 5.1.3. Identity of Hug2Go (chair + mobility vs. mobility with chair) 5.1.4. Opportunities for improvements 5.2. Limitation 5.2.1. Multi-class test 5.2.2. Controllability of steering 5.2.3. Incomplete implementation 5.2.4. Vacancy of sharing model and benefit of sharing 6 CONCLUSION 6.1. Conclusion 6.2. Further works 6.2.1. Optimizing System 6.2.2. Validation of Hug2Go indoor area 6.2.3. Interaction between driver and pedestrian 6.2.4. Development of services & business model for Hug2Go References 52 Appendices 55 Summary in Korean 157 Acknowledgement 159

Tables Table 1. Commercial product specification... 1 Table 2. insights from observation... 3 Table 3. Area of convention center in Korea... 6 Table 4. Use & Operation of hug steering... 12 Table 5. Functional requirements.... 13 Table 6. Hug2Go Specification... 14 Table 7. Demography of participants... 22 Table 8. Likert scale scoring... 25 Table 9. Respondents usage intention of the Hug2Go (n = 15)... 26 Table 10. The category of interview responses... 30 Table 11. The result of factor analysis... 33 Figures Figure 1. Forecast of new vehicles sales distribution in urban areas in the United States... 2 Figure 2. Future personal mobility opportunity area... 3 Figure 3. Research scope... 5 Figure 4. Previous study trend in PMVs industry... 7 Figure 5. Autonomous and shared... 1 Figure 6. Experimental task procedure... 2 Figure 7. Top view of driving task pathway... 2 Figure 8. Acceptance of commercial PMVs to the indoor... 4 Figure 9. PMVs Driving Environments... 6 Figure 10. The area of convention center... 7 Figure 11. Frustrated walking in huge space... 7 Figure 12. Finding passenger through self-driving inside an airport... 8 Figure 13. (a) Chair mode, (b) Self-driving mode, (c) Control mode... 9 Figure 14. Detection of position (a) Chair mode, (b) Control mode... 10 Figure 15. Passenger s hug steering posture... 11 Figure 16. The hug steering motion... 12 Figure 17. The main structure of operation... 13 Figure 18. Hardware overview. highlighting means development of level 1 of (manual mode). Not highlighting means development of level 2 (self-driving mode).... 14 Figure 19. The base platform, (a) Load cell, (b) BLDC motor, (c) Seat, (d) castors... 15 Figure 20. BLDC motor (48V, 350W), Battery (36V, 10Ah)... 15 Figure 21. Load Cell 4 wire (50kg, 2mv /V)... 17 Figure 22. Relevant position of load cell... 17 Figure 23. The process of making seat frame, (a) prepare mold, (b) spray resin, (c) laying carbon fiber, (d) finishing parts... 18 iii

Figure 24. Block Diagram... 19 Figure 25. Usability test plan dashboard... 21 Figure 26. Test platform (level 1)... 22 Figure 27. The process of usability test... 24 Figure 28. The pathway of driving test... 24 Figure 29. Manual driving in the usability test... 25 Figure 30. Acceptance of Hug2Go (n=15)... 27 Figure 31. Acceptance of Hug2Go (n = 15)... 28 Figure 32. A various types of motion... 35 Figure 33. Characteristic of product, (a) Chair, (b) Mobility... 38 Figure 34. Opportunities for improvements in the attributes of product... 39 Figure 35. The example of Hug2Go service... 44 iv

1 INTRODUCTION 1.1. Background 1.2. Research Aim and Method 1.3. Thesis structure 1

1 INTRODUCTION 1.1. Background The personal mobility vehicles (PMVs) emerged as a new category of transportation device in the early 2000s. PMVs offers many intriguing possibilities for extending the human range of mobility from about 1km to 10km or more. PMVs offers several potential benefits to consumers and society. Many researchers focused on performance or acceptability of use. However, most of PMVs regarded as outdoor mobility. Very few studies address indoor mobility. The indoor space is the daily living environment. People frequently use the airport, shopping mall, transportation. Indoor mobility also can be the area of PMVs. If shared and autonomous vehicles are adopted as quickly as other technologies (like smartphones, cellphones, and the Internet), significant variance will begin within five years and that the market for individual mobility could modify dramatically over the next 25 years (Figure 1). Recently, Mobilityon-Demand(Mod) services, such as car sharing or on-demand taxi service, have seen massive growth in the last few years through service providers like Uber and Lyft (Andersen,2016). Figure 1. Forecast of new vehicles sales distribution in urban areas in the United States In addition, many manufactures produce low-cost PMVs for leisure. For example, Segway, e- scooter, e-bike, likewise PMVs become popular for general users. PMVs is more being personalized and owned. PMVs has several potential benefits to consumers, reduced trip times, lower transportation costs. However, the full potential of the category has not been realized, because not yet light enough, do not go far enough, and cost too much for someone. 2

Recently, popular PMVs has been moved to sharing service area. For example, KICKGOING, GOGO-SSING, deer, SWING, WIND, etc. More than ten company provides on-demand service in Korea. Even though low-cost PMVs is provided for general customers, many people prefer to share than owned. It is growing up. In the transportation industry, many researcher and specialist reveal future mobility trend as personal, shared, autonomous (Scott, 2016). Through the above examples and diagram, we can figure out how it changes. The phenomenon is approaching future trend rapidly than we thought. Therefore, we need to anticipate beyond future trend and present phenomenon that we found. We thought it opportunity area for the mobility market. In this research, we suppose a new model of indoor mobility and examine it possible to build on the market through the usability test. Figure 2. Future personal mobility opportunity area 3

1.2. Research Aim and Scope The main purpose is to develop suitable indoor personal mobility concept through design-driven research. The main research question was formulated as follows: What can be done indoor space to facilitate breakthrough outdoor PMVs? Related questions derived from the main question include the following: What are the major barrier to inner space to apply existing mobility? What is the new design of indoor personal mobility? How is the acceptance of indoor PMVs? What are the implications for the indoor PMVs? 1.2.1. Research Aim Through literature review and user observation with PMVs users, we developed an indoor PMVs providing comfortable and safe driving at the inner space. Then, we evaluate the use of steering and driving through a usability test. First, Discover the opportunity for the indoor mobility Second, Design and development of a new form of the indoor mobility Third, Focused experiments and user surveys Finally, we discussed issues of results related to usability test and were able to understand the limitation and challenge of the Hug2Go design. Eventually, we reveal further works and plans. 4

1.2.3. Research Scope In this research, we present the development of indoor PMVs. The main scope of this paper are: 1) Design and development of a new indoor PMVs capable for manual control: we suggest a design form of the indoor personal mobility. 2) Embedded system has implemented: we consider building cost-effectiveness commercial product on the market. In order to lower cost, we develop a system by ourselves. 3) Usability test and user survey: we examine the intended the hug steering uses of the Hug2Go Figure 3. Research scope 1.3. Thesis Structure This paper is organized as follows. In Chapter two, a brief overview of the previous studies related to PMVs will be reviewed. In Chapter three, we describe the Hug2Go design and propose a method of manual control for the Hug2Go using the hug steering. In Chapter four, we address the procedure of a usability test. The results will be presented through an analysis of both qualitative and quantitative data. In Chapter five, the intended uses of the Hug2Go are examined using the empirical results will be discussed. In Chapter six, we present the outcomes and contributions of the research and suggest further works and directions. 5

2 PRELIMINARY STUDY 2.1 Related works 2.2 Observation to insights 6

2 PRELIMINARY STUDY 2.1. Related Works The personal electric vehicle(pev) emerged as a new category of transportation device in the late 1990s. Some studies have been conducted to analyze and discuss PMV use, especially for selfbalancing personal transporters, such as Segway. (Ulrich, 2005) pointed out that PMVs offer several potential benefits for users and society, including lower transportation costs, reduced trip times, and low environmental impact. (Sawatzky, 2007) studied the use of the Segway as an alternative mobility device for people with disabilities and concluded that subjects with disabilities through the Segway was easy to use, and we are excited about its potential to assist them. (Miller, 2008) analyzed the approach speed and passing clearance that Segway devices exhibit on encountering a variety of obstacles on the sidewalk. Figure 4. Previous study trend in PMVs industry 7

Recently, transportation and robot researchers have been increasingly interested in autonomous and shared vehicles related to PMVs. (Ando, 2013) suggests that a critical factor in successfully introducing PMVs for use in the future is understanding social acceptance. (Fujikawa, 2012) propose for an IR system to support automatic control for PMVs (Figure 5, Fujikawa 2012). They designed four-wheeled mobile bodies that are widely used in practical locomotive machinery. It is not only for outdoors but also for indoor mobility environments such as station or open public area. With multisensor based or self-driving module, some research includes perception of users and identifying open concept to the autonomous personal mobility device (Abdur-Rahim, 2016). Nevertheless, the form or design of indoor mobility seems to be lacking. The overall researchdriven system is described in Figure 5. Typicality and novelty have often been shown to be related to the aesthetic preference of human artifacts (Hekkert, 2003). The core of all question is the most advanced yet acceptable system in the design perspectives. In the next chapter, we discover the problem when the outdoor types of PMVs come into indoor space through observation. In addition, we also find insights from observation. Figure 5. Autonomous and shared 8

2.2. Observation to insights The purpose of observation is to discover the context of indoor mobility with existing PMVs driving. In order to gather observation, we conducted driving experiment with 11 people aged from 20 to 50 years old (3 females, 8 males). We use three different types of PMVs such as Segway, E-scooter (stand-on) and E-scooter (sit-on) (Table 1). Table 1. Commercial product specification Category Segway E-scooter (stand-on) E-scooter (sit-on) Product XIAOMI Nano Pro, SPEEDWAY ECORO s50 Dimension 260 x 552 x 630 980 x 380 x 1040 450 x 890 x 1090 Top speed 18km/h 28km/h 25km/h Empty weight 13.4kg 8,1kg 25.2kg Maximum payload 100kg 100kg 150kg Range 30km 20km 35km Power 63V/310Wh, 400W 36V/6.4Ah, 250W 36V/10.4Ah, 540W Charge time 3.5hr 4hr 4.5hr 9

2.2.1. Experimental Task Figure 6. displays a timeline view of the experiment. Participants required online google survey. Following, participants were informed about task. First, the experimenter gave each participant an explanation of the driving course and how to manipulate PMVs. After each of the three vehicle driving trials, questionnaires were administered to determine the participant s experience feelings (Appendix A). Finally, we interviewed each participant after finished. Figure 6. Experimental task procedure 2.2.2. Pathway Figure 7. describes experimental driving task pathway. Participants drove PMVs manually for loop course starting point to end. The participant is intended to avoid obstacles such as pillars. Figure 7. Top view of driving task pathway 10

Table 2. insights from observation Observation Insights Function As soon as using the acceleration button, She is afraid of the rapid acceleration. Many of them didn't find the power button Emergency situation He boarded the Segway without turning on power He feel fear to get out of the Segway I want to know how fast speed is e-scooter (sitting) is lower than standing type A participant wave to say hello someone on driving unconsciously He put his feet on the ground on driving She just try to push unknown button to start it A heart sound is getting more faster on video because he is nervous The space is limited when participants are turning on the pillar It is not for elderly or kids Pedestrians is latent risk The long form of mobility has risk on turning Even though they use it in indoor, they thought that it is just outdoor mobility The acceleration of speed has been limited at the starting point. Users can notice the position of power supplier. Users should control the emergency situation without difficulty Users are able to recognize controller without difficulty The indoor mobility keep users safe from unexpected situation on getting out the mobility. Users want to be aware of the speed in the driving. PMVs must keep providing user's sight. PMVs must prevent the accident from unexpected situation Users want to control the mobility completely The consequence of control should be expected Users should be aware of convenience and trust during the driving Indoor PMVs should consider the limited space. Self-balancing PMVs is not universal for everyone Indoor PMVs must consider the pedestrians in the path Users are aware of the inconvenience and risk on turning The indoor mobility should have an optimal form Safety Convenience Convenience Convenience, Safety Safety Safety, Convenience Safety, Convenience Safety Safety Safety Safety Safety Convenience Safety Safety Appearance Sitting is safe PMVs should be comfortable on driving Safety Segway enable us to be hands-free User want to hands free Convenience If we meet the shared mobility, we have to spend time to be used to mobility The space is limited in door (elevator, corridors, etc.) The indoor mobility should provide trust to users in sequence of driving (boarding, getting out, driving) The indoor mobility should have an optimal size Convenience Appearance 11

2.2.3. Insights and findings In short, we obtained twenty-three of insights from observation (Table 2). Then, twenty-three of insights can be connected with functional requirements. Thirteen of insights are related to safety; eight of insights are the convenience; others are appearances. We conducted a pre-google survey for gathering quantitative data. The experimental results (Figure 8) show that the value of acceptance, such as safe, fun, intuitive, universally accessible, efficient, fun is entirely positive. However, there is an exception to the result. Segway is not enough. It follows from what has been said that we should consider insights from commercial PMVs and we figured out the possibilities of the indoor mobility. Figure 8. Acceptance of commercial PMVs to the indoor 12

3 Hug2Go The Indoor Smart Driving Personal Mobility 3.1. Huge indoor spaces 3.2. Design features 3.2. Implementation 13

3 Hug2Go 3.1. Defining target area We define the target space as it is the large indoor spaces segment. It includes building which has a huge indoor area such as malls, station, museums, convention center and airport (Figure 9). A convention center will be the target space. It is a modern convention center in which one or more buildings from a complex of shops with interconnecting walkways. Others such as station, museums, convention center and airport also have huge indoor space with pathways. Probably, people are walking along passage connecting different sections of a building in these malls. It frequently happens in the huge indoor environment. Figure 9. PMVs Driving Environments Specifically, the area of huge indoor spaces is more than 20,000m 2 representative convention center in Korea (Table 3). at least. There are Table 3. Area of convention center in Korea EXCO COEX BEXCO KINTEX Location Daegu Seoul Busan Goyang Area (m 2 ) 26,508 36,736 54,731 108,483 14

Figure 10. The area of convention center Figure 10. show the area of convention center. The walking makes people exhaust in a large space. Even though they want to spend going somewhere, it s physically exhausting. It s a long journey in large space. We can assume the traveling in convention center. Figure 11 describes comparison between the average daily walking and traveling in convention center. Approximately, Traveling is ten times than the average of daily walking distance. It s obviously long journey for walking. Therefore, it is certain latent needs and we are targeting on a large space. Figure 11. Frustrated walking in huge space 15

3.2. Design features Hug2Go is the indoor personal mobility, finding passenger through self-driving and going to place by a new way of steering. The mobility is intended to work corporately with people to improve the efficiency of time or energy indoor environment. Primarily, Hug2Go provides three types of operating mode. It usually requires self-driving for finding the passenger. Sometimes, it could be a comfortable chair. If a user wants to control the mobility, they are also able to operate the steering. Therefore, the switching is essential part among mode use. In order to discern between chair mode and control mode, we use sensors to detect a passenger s position. Finally, we suggest hug steering as a new way of steering. we called it hug steering because a user needs to hug seat back to control the mobility. It might be a quite new steering system. Figure 12. Finding passenger through self-driving inside an airport 16

3.2.1. Three types of operating mode Figure 13. (a) Chair mode, (b) Self-driving mode, (c) Control mode Hug2Go provides three types of operating mode (Figure 13, a, b, c). The modes can be changed by each other. First, chair mode supports temporary breaks for users. The passenger can sit and take a break comfortably when the mobility is stopped. Second, a self-driving mode is finding the passenger who wants to ride on the mobility. Hug2Go maintains a safe distance with people through multiple sensors while the self-driving mode is moving. When users send gestures or voices for boarding signs to mobility, Hug2Go will reach over the passenger. Third, the control mode means manual operating. If the passenger wants to control the mobility, they can use facing seat back as steering. Users should sit on a seat back in the opposite direction. Facing back becomes steering. They can control the direction through the rotation, push, and pull. As above mentioned, Hug2go enable a passenger to select different modes. 4 ToF (Time of Flight) sensors are installed to detect passenger s position. The overall location of sensors is described in Figure 14. The sensors are mounted on seat and front cover where they detect passenger s leg or part of the body. 17

3.2.2. Switching modes Figure 14. Detection of position (a) Chair mode, (b) Control mode 3.2.3. Sitting on chair (chair mode) On self-driving, the mobility reaches to passenger and stops in front of the user. Unless the user wants to control the mobility, the user can sit on leaning back. Then, 2 ToF sensors detect a user s part of the body under the knee. If ToF sensors detect user s body, Hug2Go is not controlled by any forces. (Figure 14, a) 3.2.4. Sitting on a chair facing in the opposite direction (control mode) In order to control the mobility, a user must sit on a chair facing on the opposite side. Then, 2 ToF sensors which is mounted on front cover detect passenger s leg or part of the body. If ToF sensors detect the user s part of the body, Hug2Go s control mode is only activated. After that, the user can control the mobility through using hug steering.(figure 14,b) 18

3.2.5. The hug steering Steering of mobility is an essential activity. However, the most traditional steering is still insufficient. Especially, there is no optimal standard related to indoor personal mobility. We needed a new way of steering suitable for indoor mobility. In this paper, we follow the form of the chair. Therefore, we consider relevant steering for Hug2Go as soon as possible. The most basic ideas of criterion are from sitting on a chair facing in the opposite direction. Everyone might have an experience such an opposite posture in the chair. This is not stable but pleasant. We believe that such an experience makes it funnier steering experience. Finally, we can suppose hug steering (Figure 15). The hug steering is steering implemented in Hug2Go. It is a new way of steering, which means that it contains a hug motion. Hug is squeezing someone tightly in one s arms, or holding something carefully or tightly round or against part of one s body. Figure 15. Passenger s hug steering posture 19

Figure 16. The hug steering motion The hug steering produces four types of steering motion (Figure 16). User can push, pull, and rotate hug steering. If the user push the seat back, the mobility is moving forward (Figure 16, a, b). To pull the steering means moving back. In order to turn in a circle, the user rotates the seat back to left or right without pull and push. Then, the mobility is rotating toward rotating direction (Figure 16, c, d). The user probably needs an operating rotation of mobility on moving forward or back. It is a different operation with turning in a circle when it stops. If the user pulls or pushes and rotates the steering simultaneously, the mobility moves the direction to which the user rotates on driving. The strength of the rotating makes how it moves quickly. Table 4. Use & Operation of hug steering Types Use Operation Direction (a) Push Move forward (b) Pull Move back (c) Rotate (right) Turn right (d) Rotate (left) Turn left 20

3.3. Implementation In this research, we focused on manual mode. Hug2Go two kinds of operation (Figure 17). The self-driving is an essential part of our development. The development of autonomous personal mobility devices has been an active research area recently. Many researchers are developing autonomous mobility. However, It still costs huge money. One of the focus in this research is to provide a lower cost platform than previous research platform. Therefore, we examine the possibility of low cost and plan to develop self-driving mobility step by step. It will develop in the next research. Figure 17. The main structure of operation First, we begin to prepare functional requirements (Table 5) in order to develop manual mode. The typical architecture of a personal electric vehicle is comprised of the essential functions of energy storage, drive system, and chassis. We discuss the driving system and chassis without self-driving in this research. Table 5. Functional requirements. Mode Function Component Part Moving, rotating Motor BLDC motor Manual Selfdriving Control Steering Load cell Sitting / mode change Seat Seat Mapping & localization Sensor ToF sensor Self-charging Charging Battery, auto-charging system Obstacle detection Sensor ToF sensor 21

3.3.1. Hardware Table 6. Hug2Go Specification Dimension () Empty Weight Maximum payload Maximum speed 80cm x 60cm x 120cm 25kg 100kg 8.0 km/h Range 4.5h Power 36V, 10Ah Figure 18. Hardware overview. highlighting means development of level 1 of (manual mode). Not highlighting means development of level 2 (self-driving mode). The overall design describes a round plane figure whose boundary consists of points equidistant from a fixed point. The platform consists of six parts (Figure 18). In this research, we focused on manual mode. Therefore, we developed a seat, steering, and motors. A metal structure customizes the base platform. It is the main body which connects with a seat back, seat, wheel, and canisters (Figure 19). t draws the shape of the ellipse. Both edges place to seat and seat back. The motor is mounted to under metal structure. Steering actuation is achieved using two load cells. We will deal with details below. 22

Figure 19. The base platform, (a) Load cell, (b) BLDC motor, (c) Seat, (d) castors 3.3.2. Motor Two BLDC motors are installed to move the mobility s driving wheel. The mobility comes with 36V, 10Ah internal lithium-polymer battery, used to power the main motor as well as the accompanying primary circuit. BLDC motors offer advantages over brushed DC motors, including higher reliability, longer lifetime (no brush erosion), elimination of ionizing sparks from the commuter, and the overall reduction of electromagnetic interference. Brushless motors are considered more efficient than brushed DC motors. This means for the same input power. A brushless motor will convert more electrical power into mechanical power than a brushed motor. Figure 20. BLDC motor (48V, 350W), Battery (36V, 10Ah) 23

3.3.3. Steering Manual controls of the mobility can only be achieved through output signals with two load cell installed. Additionally, stop-button is also used to stop it when the vehicle is out of order from the intended path. It prevents a potential safety hazard. The longitudinal speed control of the vehicles is achieved by an analog voltage input to the motor driver to imitate the output from independent two load cells linked to the main body. There is two control input to the mobility: steering and speed. The speed input value to the steering can also take negative value indicating reverse motion. The central controller channel takes an input signal of 0-3.3V, with 0V being zero speed, and 3.3V being maximum forward speed. The maximum velocity of the mobility is set limited to 5km/h. In order to operate manual driving, the vehicle needs two beam types of the load cell. A load cell is 130mm 30mm 22mm (). The rated max output is 2mV/V. It is the output voltage when the rated capacity of 50kg is loaded. A load cell consists of a metal element that is introduced to a change through tension (pulling apart) or compression (pushing together) forces and interior strain gages that sense this change. Then, the hug steering needs four types of motion. Therefore, the load cell provides independent four signals. Normally, a load cell makes two signals at least through tension and compression. We used two load cell to make four signals from the independent position of the load cell. Proper position of the load cell is described in Figure 22. Two load cell is mounted to the frame covered with a seat back. It means that mounted two load cell is connected indirectly. Therefore, we need to determine the relevant position through the test. Finally, we selected the position (Figure 22, c, d) because of the independent relationship between each load cell. 24

Figure 21. Load Cell 4 wire (50kg, 2mv /V) Figure 22. Relevant position of load cell 25

3.3.4. Seat and seat back The seat and seat back enable users to rest comfortably. We regarded seat as a critical factor of safety. We considered using a soft foam pad. In order to assemble a cushion and back frame, we need to put the cover on the foam pad and assemble the rear frame to assemble a basic seat shape. In this research, we only want to focus on frame parts because we did not decide the shape of foam and choose material yet. Figure 23 describes the process of making a seat frame. We selected carbon fiber as materials. Carbon fibers have several advantages, including high stiffness, high tensile strength, low weight, high chemical resistance, high-temperature tolerance, and low thermal expansion. First, we prepared a mold for combining polyester resins (Figure 23, a). CNC machines make it excellent shape from 3d modeling. Typically, polyester resins reinforce mold more harden (Figure 23, b). After filled with resin, we can bond together with fabric. The fabric is composed of woven carbon filaments (Figure 23, c) The bonded surfaces between resin and fabric take about 12 hours. Finally, we apply at least three more layers of resin. These layers of resin are intended to improve the look of the part, not the strength or functionality. After that, we sand the top layer of resin and polish the part (Figure 23, d). Figure 23. The process of making seat frame, (a) prepare mold, (b) spray resin, (c) laying carbon fiber, (d) finishing parts 26

3.3.5. Embedded system Figure 24. Block Diagram An STM32L432KC microcontroller is used to publish a control signal to the BLDC motor, to read the load cells and communicate with the hug steering. we use Arm MBED OS, which is an opensource embedded operating system. It includes all the features to develop connected product base on Arm Cortex-M4 at 72MHz microcontroller, including security, connectivity, and RTOS. A customdesigned circuit board is mounted to mobility inside. The passenger of mobility has access to the board on mobility. The ToF sensors detect the signal whether the passenger has arrived. In order to detect the passenger s motion, it needs two at least. There are two available modes: chair, manual. In manual mode, only commands from the steering will be executed; in char mode, the mobility will stay stationary. 27

4 USABILITY EVALUATION 4.1. Aim & Plan 4.2. Method 4.3. Results 4.4. Findings 28

4 USABILITY EVALUATION 4.1. Aim & Plan We conducted a usability test using a working prototype to evaluate the validation of the Hug2Go concept in terms of safety, comfort, ease of use. In order to avoid unnecessary misunderstanding, we developed the usability test plan toolkit (Figure 25). It is a detailed document that describes everything from getting to the test venue to the exact words the test moderator will use. Through the usability evaluation, we aimed to ask as followings: Do people understand the value of Hug2Go? Do people understand how to select different kinds of modes? Do people understand how to use the hug steering? Figure 25. Usability test plan dashboard 29

4.2. Method In order to usability evaluation, we used below test platform (Figure 26). In particular, we conducted MVP (minimum viable product), which is to provide feedback for product development. The seat and seat back are only designed and developed with the motor system to be verified from initial users. Figure 26. Test platform (level 1) 4.2.1. Participants The usability test was conducted with 15 people from 20 to 30 years old (8 males, 7 female). According to (Nielson, 1993), At least, 15 users are needed to discover all the usability problems in the design. It is probably possible to cover 100% of usability problems. Therefore, we recruited 15 peoples. The demography of the 15 participants are presented in Table 7 Overall, 53.3% of the participants were male, and 46.6% were female. The study group consisted of the followings: 15 persons (100%) are with ages 20-30. Overall, 33.3% of the participants have no experience of PMVs. 60% of the participants have a few times of experience. 0.6% of the participant only has lots of experience. Participants entirely are not aware of the existence of PMVs still. 30

Table 7. Demography of participants Participants Gender Age Experience of PMVs P1 Male 20-30 Nothing P2 Male 20-30 Nothing P3 Male 20-30 Nothing P4 Male 20-30 A few times P5 Male 20-30 A few times P6 Male 20-30 A few times P7 Male 20-30 A few times P8 Male 20-30 A few times P9 female 20-30 Nothing P10 female 20-30 A few times P11 female 20-30 Nothing P12 female 20-30 A number of experience P13 female 20-30 A few times P14 female 20-30 A few times P15 female 20-30 A few times 31

4.2.2. Experimental procedure Figure 27. displays a timeline view of the usability test. Participants required to be informed about the task before the usability test. First, the facilitator previewed experiment explanation. Particularly, participants were able to understand the main concept of Hug2Go. Second, participants spent enough time to practice driving. We demonstrated how to use the hug steering before manual trials for participants. Third, participants started to operate Hug2Go by themselves. Then, we gave the essential and formative information related to steering. Participant carried out manual trial. They move along the driving route (Figure 28). Previously, we consider the pathway where participant drives safely. Finally, participants take in a survey concerning the acceptance of Hug2Go and discuss it through the interview. Figure 27. The process of usability test Above mentioned in Chapter 3.1, we are targeting huge indoor space. Therefore, the 1F lobby, which is located in UNIST (approximately the area of 3600m2 )is regarded as enough huge indoor space (Figure 9). Figure 28. The pathway of driving test 32

Figure 29. Manual driving in the usability test 4.2.3. Experimental design The user evaluation measured the participant s perception of the Hug2Go in terms of safety, comfort, ease of use. Comfort was defined as the quality of ride and seated well-being: Safety was defined as the user s level of control of the hug steering and his/her feeling of stability when driving the Hug2Go. We referred to a driving assessment tool to conduct user evaluation in terms of indoor personal mobility. The PMCDA developed by (Karamaj, 2014) was a validated training method for a power wheelchair. The PMCDA tool was used by a clinician who assessed the participant s driving skills in a scale 1 to 3 in the area of adequacy-efficacy (AE) and safety. The questionnaire (Appendix C) was modified to fit the tasks of the indoor spaces through PMCDA (Appendix B). Each measurement used a Likert scale of 1-5 where a score of 1-strongly disagree, 2-disagree, 3-neutral,4- agree, 5-strongly agree (Table 8). he total score was collected by the session of the questionnaire. Table 8. Likert scale scoring Strongly Disagree Disagree Neutral Agree Strongly Agree ( ) () ( ) () ( ) 1 2 3 4 5 33

4.3. Results 4.3.1. Usage Intention of Hug2Go The summary of usage intention of Hug2Go is in Table 9. Among the respondents, 15 respondents (100%) preferred the Hug2Go for moving within a building. Moreover, eight respondents (53.5%) believed that the Hug2Go could be used by disabled or.35) older people. Furthermore, five respondents (33.3%) preferred the Hug2Go for accessibility between home and nearest station and going shopping in the neighborhood. On the other hand, only one respondent (6.6%) thought the Hug2Go was available to go business travel in the urban area. Additionally, no respondent believes that medium-distance commute or daily transport in an urban area. The survey results show in Table 9. indicate significantly valid for indoor mobility. All respondents preferred for moving within a building. However, 20-30 ages of the respondents only participated in this usability test. After recruiting multi-class people, this study is needed for validation. Table 9. Respondents usage intention of the Hug2Go (n = 15) Usage Intention Frequency Percentage (%) Moving within a building 15 100 Transport support for disabled or elderly people 8 53.3 Access between home and the nearest station 5 33.3 Shopping in neighborhood 5 33.3 Short-distance trip in downtown area 4 26.6 Extended travel ranges as a substitute 4 26.6 Access between destination and train station/ bus stop 3 20 Going to the neighborhood hospital 3 20 Touring and excursions 2 13.3 Business travel in urban area 1 6.6 Medium-distance commute 0 0 Daily transport in urban area 0 0 *The respondent is available for multiple selections 34

4.3.2. Acceptance of the indoor PMVs Factors Mean Pleasant 4.3 0.88 Desirable 4.1 0.61 Useful 3.7 0.97 Likeable 3.6 0.82 Practical 3.6 0.73 Convenient 3.5 0.35 Effective 3.4 0.99 Figure 30. Acceptance of Hug2Go (n=15) The seven factors used to assess the respondent s judgment. The significant logical factors can influence acceptance of PMVs (Ando 2013). The value is regarded as the average of scoring (n=15, Likert scale of 1-5). As mentioned Sect. 4.2.3. The total score was used a Likert scale of 1-5 (Appendix B. ). Figure 30 shows the result in terms of acceptance. Overall, the seven factors are more than the average value of 3. It indicates that respondents are impressed by Hug2Go as acceptable indoor mobility entirely. In particular, Pleasant and Desirable are highly scored in these factors. On the other hand, Convenient and Effective are lower compared with other factors. 35

4.3.3. Usability of manual driving We aimed at evaluating the hug steering through a usability test of manual driving. Before the usability test, we encouraged participants to drive the Hug2Go by themselves. During the manual driving (1,2 loops), participant evaluates the usability of the Hug2Go. Appendix D. shows that the average of the score is lower than score four overall. It indicates that participants were not satisfied with the manual driving remarkably. The lack of satisfying response could be due to the control system is not optimized. On the other hand, the usefulness of the indoor is adequate for participants. Furthermore, Q 3.11, Q 3.3, Q 3.8, Q 3.10 show the score of more than 3 (Appendix D. The mean of usability test (n=15, Maximum score 5)). Figure 31 describes the acceptance of Hug2go from Appendix D. It shows three kinds of responses: disagree, neutral and agree. Figure 31. Acceptance of Hug2Go (n = 15) 36

4.3.4. Interview We used interviews to collect rich insights. The interviews were initiated by asking several questions in terms of usability of the hug steering and acceptance of the concept. After that, we discussed the major problems and pains. The interview is a 20 to 30 min, semi-structured interview that assesses the Hug2Go users participation in activities. Besides, we conducted all the interviews face-to-face individually. The category for this interview consisted of five types of entities: the hug steering, seat, and seat back, safety, aesthetic, others. They represented control, design, functional requirements. After completing the 15 interviews, the research team transcribed and translated the voice recordings. We constructed the significant findings based on classification the process of transcribing, translating resulted in the identification of 5 entities (Table 10). First, many of interview response reveals the difficulty of operation. In particular, when participants meet the first impression, many participants are confused about identity weather chair or mobility. Second, the participants did not sit back fully, leaning on the seat back. It is a different result expected. They want seat soft and more comfortable. Third, speed is suitable for indoor mobility. The participant mentioned that a reason would be due to the accident with a pedestrian caused by fast speed. Also, they want to stop in the emergency. Fourth, a more familiar design is needed for harmony with the indoor environment. Finally, one of the participants suggests the stop station rather than selfdriving. In details, as following: 1) The hug steering 2) Seat and seat back 3) Safety 4) Aesthetic 5) Others 37

Table 10. The category of interview responses Category Interview responses The hug steering P01 P01 P04 P05 P06 P06 P06 P07 P08 I was confused as to whether it was a seatback or a controller. I thought of it as a steering wheel. I hope it s easy to go forward and turn left or right simultaneously. I didn t expect to use it before I see it. I was surprised that the seat back is a controller If I keep turning in the floor, I feel dizzy. The chair is hard, so I feel uncomfortable. I need significant physical load when I control the hug steer. First boarding the control is not as easy as I though before. Second boarding it s fun to learn how to operate It is difficult unless you explain the way of control. It is a lack of affordance. I hope the feeling of hug as being inside come to volume. It s easy to learn but It is difficult to recognize how much force to turn. P13 It s difficult to operate it. P01 P01 P01 P03 It was uncomfortable to sit with your back full It was uncomfortable to sit with seat back. I m afraid of moving when I sit down. If you had a handle or something, it would be better. I wouldn t have to look back when I went back. Hard seat, It is irritating on pedestrian s eyes P04 The usage was not easy. It was far from the chair. Seat and Seat back P05 P07 P08 P08 P10 I don t know where to sit. Feel uneasy sitting down when I sit I thought the curves would be good for sitting. But It was slipped. When I drive, the foot rest were convenient. It s inconvenient to sit on the floor. You don t have to use your seat back I hope the seat is soft. I felt like sitting on a structure. It would be better if it was leather seat. the process of use is not difficult. There is an uncomfortable side to sit back. P10 It is hard for women to get on board because of their skirts. 38

P11 P11 The chair is hard. Difficult to operate, the seat back is not intuitive. It seems an inappropriate mode of operation for women. The armrest is needed. P12 The seat back is too far P13 P05 P05 I hope it s more soft. It s very slippery. It was good to move slowly in the indoor space. I want to keep riding because I like it. It was safe braking because I can get off whenever I m in dangerous situation. P12 It s dangerous if your knees come forward. Safety P04 P05 I wish stop button or seat handle. I trust power on/off button. I hope familiar and safe design. P08 If I moved back, it would sound P10 Pedestrian want to know the location of mobility through indication of lights or sound. It will be better to stop near the pedestrian even though you didn t perceive the pedestrian on driving. P06 It s scary because of the steel frame. Aesthetic P11 It new but unfamiliar. The part of carbon is not suitable. Others P13 P07 you need to choose bright color. It must not be a wheelchair design. If you considered the curve design of seat rather than the cushion. It is better to the stop station. I hope there was container for load. I need an emergency brake button. 39

4.4. Findings 4.4.1. Valid PMVs for the indoor space Interestingly, a hundred percentage participant agrees with the acceptability of indoor mobility (Table 9). Moreover, among the respondents, 13 respondents (86.6%) responded that it is useful to drive to the indoor space (Appendix D). Regardless of the low fidelity of the platform, most of the respondent thought that it is highly validated for the indoor space. 4.4.2. Incomplete manual steering structure Eleven respondents (73.3%) responded that it is not easy to use without the user guide. Furthermore, eight respondents (53.3%) disagreed that it is easy to control moving forward, back, left, and right. On the other hand, P08 mentioned, it is easy to learn and a half of respondents felt that it is intuitive to control the mobility (Appendix D). Therefore, hug steering is somewhat unclear. It could be due to incomplete manual steering structure. 4.4.3. Not easy but pleasant Appendix D. shows the significantly counted negative feedback is to control the hug steering. As mentioned in 4.4.2, It depends on the fidelity of the hug steering. However, generally, the participants enjoyed the Hug2Go, as evidenced from many indicating that they would want to ride the Hug2Go again in the future and comments on surveys such as fun or awesome. 40

4.4.4. Correlation component with acceptance through Factor analysis For the acceptance of the indoor PMVs, factor analysis was used to describe variability among observed, correlated variables in terms of a potentially. The goal of this analysis was to build a correlation between the factors. The seven-factor used to assess the respondents judgments about the PMVs. A Kaiser normalization was performed to define the components in terms of subsets, as shown in Figure 30. The first subscales, termed practical, useful, effective could be interpreted as the Base1 of a system. The second set of subscales, termed convenient, likable, could be interpreted as reflecting the Base2 with a system. The third is pleasant. The fourth is desirable. On the other hand, in Appendix D the mean scores for practical and effective are much less than others. It indicates that the test platform provides enough usefulness to users. Table 11. The result of factor analysis Base1 Base2 Base3 Base4 Practical 0.949 0.076-0.017-0.006 Useful 0.916 0.060-0.097 0.176 Effective 0.816 0.299 0.204 0.248 Convenient 0.007 0.974-0.053 0.022 Likable 0.423 0.808 0.274-0.103 Pleasant -0.003 0.071 0.983-0.108 Desirable 0.192-0.040-0.116 0.968 Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. a. Rotation converged in 5 iterations. 41

4.4.5. Insights through interview After completing the 15 interviews, the research members reviewed and interpreted with the interview. We constructed significant insights based on sorting. The process of transcribing, translating resulted in the identification of several categories as followings: 4.4.6. Recognition of moving reverse The participant is not used to operating the hug steering yet. They are used to operating existing mobility. A participant did not recognize reversing function by the end. It could be because of no reverse function in the existing PMVs. 4.4.7. A lack of the steering feedback The hug steering is no feedback with a level of control. It only depends on the driver s control. Typically, the controller has such an indicator. For example, throttle or joystick enables users to recognize the degree of control through physical feedback. P08 commented, easy to learn, but It is difficult to recognize how much force to turn. 4.4.8. Emergency button and pedestrian verification The Hug2Go has a function of reverse moving. During the reverse moving, a passenger could not recognize the pedestrians. Therefore, we consider the interaction between pedestrian and passenger. Also, in order to avoid an emergency, users want to stop immediately. P04 commented, I wish stop button or seat handle. and P05 said, I trust power on/off button. I hope familiar and safe design. P8 and P10 said, If I moved back, it would sound and Pedestrian want to know the location of mobility through the indication of lights or sound. It will be better to stop near the pedestrian even though you did not perceive the pedestrian on driving. More, the participants also reported that the highest levels of discomfort and fear without the assistance system ate the high pedestrian density. 42

4.4.9. Uncertain affordance (Hug vs. Hold) We observed a various type of motion to control the hug steering. It indicates uncertain affordance of steering control. P01 addressed, I was confused as to whether it was a seat back or a controller. I thought of it as a steering wheel. P04 described, I did not expect to use it before I see it. I was surprised that the seat back is a controller Figure 32. A various types of motion 4.4.10. Adjusting seat and seat back height or posture Probably, the user s height or posture might affect the performance of the hug steering. It could due to the form of mobility. The Hug2Go follows general chair form. Consequently, a user s height or position will be designed parameter. P01 said, It was uncomfortable to sit with your back full and P12 said The seat back is too far. P05 commented, I do not know where to sit. Feel uneasy sitting down. 4.4.11. Enhancing convenience P07 addressed, I hope there was a container for the load. 43