Linköping’s university | Department of Computer and Information Science Master thesis, 30 credits | Master of Science - Design and Product Development Spring 2020 | LIU-IDA/LITH-EX-A--20/018--SE
Autonomous cleaning robot as a
service
Exploring value co-creation opportunities for an
autonomous cleaning robot in the context of a
cleaning service
Pernilla Sandén
Jeff Pertot
Supervisor: Ana Kustrak Korper Examiner: Johan Blomkvist
Linköpings Universitet SE-581 83 013-28 10 00, www.liu.se
II
Copyright
The publishers will keep this document online on the Internet – or its possible replacement – for a period of 25 years starting from the date of publication barring exceptional circumstances.
The online availability of the document implies permanent permission for anyone to read, to download, or to print out single copies for his/hers own use and to use it unchanged for non-commercial research and educational purpose. Subsequent transfers of copyright cannot revoke this permission. All other uses of the document are conditional upon the consent of the copyright owner. The publisher has taken technical and administrative measures to assure authenticity, security and accessibility.
According to intellectual property law the author has the right to be mentioned when his/her work is accessed as described above and to be protected against infringement.
For additional information about the Linköping University Electronic Press and its procedures for publication and for assurance of document integrity, please refer to its www home page: http://www.ep.liu.se/.
III
Abstract
The purpose of this thesis is to explore the value co-creation opportunities for an autonomous cleaning robot, in the context of a cleaning service, by using several prototyping methods. Assuming a service perspective, with an exploratory, multi-stakeholder, and collaborative approach, a case study of an autonomous cleaning robot currently in development is conducted.
With the methodological approach of sequentially and iteratively going through the phases of Mapping and Ideation, Conceptualization, Prototyping, and Data analysis, a total of 10 co-creative, stakeholder-centric workshops were conducted, using various prototyping methods e.g. Contextual value network mapping and Desktop walkthrough.
Findings show that the value that an autonomous cleaning robot can bring different stakeholders in the context of a cleaning service is multi-facetted and can be divided in to eight different Aspects, where each value carries a different implication at the level of the individual and the level of the system. Furthermore, a list of seven different Design parameters and four different Design problems are presented, which should be used as guidelines when further developing a cleaning service that uses an autonomous cleaning robot.
Finally, the implications of the findings are put into a broader perspective through discussion, with topics such as the impact of more robots in society.
IV
Acknowledgement
We would like to thank everyone that has in one way or another, made it possible for us to conduct this project. We would like to first give our thanks to the people at Dyno Robotics for allowing us to write this thesis at their company and has made it possible to conduct different activities with the help of their resources.
We like to give our most sincere thanks to our supervisor at Linköping’s University, Ana
Kustrak Korper, for your very helpful feedback. We are very grateful for the time you have
spent on giving suggestions to research topics and papers, and all the feedback regarding the thesis paper.
We would also like to thank our examination at Linköping’s University, Johan
Blomkvist, for your valuable input regarding the thesis.
Also, an extra thanks to our opponents, Gustav Backhans and Douglas Driving for the valuable feedback you have given on our work.
Lastly, we like to give a special thanks to all the stakeholders and participants that have participated in the activities we have conducted. Your input and ideas are the foundation which this thesis results are built upon.
V
Table of content
Introduction ... 1 Background ... 1 Purpose ... 1 Research questions... 1 Dyno Robotics ... 2 Limitations ... 2 Disposition ... 3 Theory ... 4 Design of a service ... 4Value and value co-creation... 5
Prototyping ... 6
Service robots ... 7
2.4.1 Definition and application ... 7
2.4.2 Introducing service robots ... 8
2.4.3 Levels of autonomy in service robots ... 9
Method ... 10
The process overview ... 10
Data gathering and insight generation ... 12
Mapping & Ideation ... 13
3.3.1 The workshop ... 13
3.3.2 Map generation and analysis of the function cards activity ... 15
Conceptualization ... 17
3.4.1 The Collaboration Robot ... 17
3.4.2 The Nighttime Robot ... 18
Prototyping ... 18
3.5.1 Desktop walkthrough ... 18
3.5.2 Experience prototype ... 19
3.5.3 Service advertisement ... 22
Synthesis of gathered data ... 24
Results ... 25
Current and future service map ... 25
4.1.1 Current service map ... 26
4.1.2 Value exchange map ... 27
4.1.3 Future service map ... 28
Aspects of value ... 28
Design parameters and design problems... 31
Discussion ... 35
Results in a larger context ... 35
Possible variations in the future cleaning service ... 35
Priority amongst the Aspects of value ... 36
Different values at different levels of automation ... 37
The accuracy of the service maps ... 37
Why a collaborative approach ... 37
Prototyping methods ... 38
Source criticism ... 38
Conclusion ... 39
Future work ... 40
VI Appendix ... 45 Appendix 1 ... 45 Appendix 2 ... 59 Appendix 3 ... 65 Appendix 4 ... 68 Appendix 5 ... 70
Figures
Figure 1: The sweeping robot that is currently in development at Dyno Robotics. ... 2Figure 2: Service delivery based on the complexity of emotional and cognitive tasks (Wirtz, o.a., 2018). ... 8
Figure 3- Service delivery based on volume and heterogeneity of physical tasks (Wirtz, o.a., 2018). 8 Figure 4: Examples of the actor cards. Icons made by Freepik from www.flaticon.com ... 13
Figure 5: An example of the current state mapping activity in progress. Icons made by Freepik from www.flaticon.com ... 14
Figure 6: An example of the future state mapping activity in progress. Notice that in this particular example, the only difference between the future and current state service map is the addition of the robot actor in the upper middle section of the picture. Icons made by Freepik from www.flaticon.com ... 14
Figure 7: Examples of function cards. Icons made by Freepik from www.flaticon.com ... 15
Figure 8: An example of the cleaning service as seen from one stakeholder ... 16
Figure 9: An example of the future cleaning service as seen from the perspective of one stakeholder ... 16
Figure 10: An example of a part of an Affinity diagram. ... 17
Figure 11: Setup for the desktop walkthrough. ... 19
Figure 12: Example of a desktop walkthrough in action. ... 19
Figure 13: The experience prototype being carried out. ... 21
Figure 14: The poster that was sent to key stakeholders. ... 23
Figure 15: Insights gathered through different phased and methods. ... 24
Figure 16: The current cleaning service map. ... 26
Figure 17: The value exchange map of the current cleaning service. ... 27
Figure 18: The future service map. ... 28
Figure 19: Illustration of the cleaning service system at property owner 1. ... 59
Figure 20: Illustration of the cleaning service system at the retailer. ... 61
Figure 21: Illustration of the cleaning service system at the outdoor cleaning company. ... 62
Figure 22: The illustrated version of the cleaning service system for the indoor cleaning at property owner 2.. ... 64
Figure 23: Illustration of the cleaning service system with an autonomous robot – Property owner 1. Figure 24: Illustration of the cleaning service system with an autonomous robot - Retailer. Figure 25: Illustration of the cleaning service system with an autonomous robot - Outdoor cleaning company. ... Figure 26: Illustration of the cleaning service system with an autonomous robot - Property owner 2. Figure 27: Large version of traditional cleaning service map ... 70
VII
Tables
Table 1 – Which aspects of the service that the different perspectives of the value co-creating system, the socio-material configuration, and the service encounter relates to and which prototyping techniques were used to explore them. ... 10 Table 2: The different phases of this thesis. ... 11 Table 3: The process of insight generation ... 12 Table 4: Example of the order of the modes of interaction for some of the participants. For every mode, scenario 1-3 was performed with the same order. ... 21 Table 5: (System Usability Scale (SUS) Plus, u.d.) ... 22 Table 6: Aspects of value. Each aspect of value carries a different meaning depending on if it is viewed from the individual/experiential level or the system level. ... 29 Table 7: Design parameters and design problems to be considered when developing and implementing a future cleaning service that uses autonomous cleaning robots. ... 31
1
Introduction
This chapter gives an introduction and background to the work of this thesis as well as the company Dyno Robotics. Research questions are introduced as well as the purpose of the thesis.
Background
Cleaning work is demanding and takes its toll on the body. With health hazards ranging from ergonomic strain to exposure to hazardous chemicals and pollutants (European Agency for Safety and Health at Work, 2020), there is good reason to augment the cleaning industry with robots that perform part of the labor. Furthermore, since cleaning work tends to be lowly paid, and have workers that are lowly engaged and thus like have low motivation, robots may even be preferred by both employers and costumers (Wirtz, et al., 2018).
Putting an autonomous cleaning robot into a cleaning service requires to some extent a redesign of certain elements of the service. However, since cleaning services can be regarded as complex services (Patrício, Fisk, Falcão e Cunha, & Constantine, 2011), redesigning the service with only costumers and employers in mind runs the risk hindering the value of the service for other stakeholders in the service. Therefore, it is important to approach the design of a new service in a holistic manner, and understand which stakeholders there are and what they respectively value.
Service design is a field of design that presents an approach for handling the complexities of designing a service (Lawrence, Schneider, Stickdorn, & Hormess, 2018). It emphasizes the collaborative, multi-stakeholder nature of services, and presents both general principles one can abide to as well as concrete methods one can use. The heart of service design is trying to find out what different stakeholders in a service value.
Purpose
The purpose of this thesis is to explore the value co-creation opportunities for an autonomous cleaning robot in a cleaning service.
Research questions
Based on the purpose this thesis will address the following research questions:
RQ1: What do current cleaning services look like, and how should a future cleaning service that uses autonomous cleaning robots look like in order to be conducive to the value co-creation opportunities found in this thesis?
RQ2: In what different ways could different stakeholders – such as cleaning companies, property owners and cleaning product retailers – value a future cleaning service that uses autonomous robots?
RQ3: Which design parameters and design problems are important to consider in the implementation of identified value co-creation opportunities in a future cleaning service that uses autonomous cleaning robots?
2
Dyno Robotics
Dyno Robotics is a startup in Linköping that specializes in developing robotics solutions. It was founded in 2018 and consists today of 6 people whose academic and professional backgrounds mostly reside within various domains of engineering, such as mechanical engineering, electrical engineering, and software development.
One of their currently active projects is the development of an autonomous sweeping robot. Figure 1 below shows the physical prototype of said robot. It uses a lidar, GPS and other technologies to autonomously navigate its surroundings. This thesis is done in collaboration with Dyno Robotics, and the different methods devised in this thesis will be applied to the development of their future cleaning service. Furthermore, parts of Dyno Robotics work, such as the physical prototype, will be used as part of some of the prototypes mentioned later in this thesis.
While Dyno Robotics is specifically developing an autonomous sweeping robot, the specific technologies that they are developing for the robot are transferable to other types of cleaning machines, e.g. a scrubbing robot. Throughout this thesis, the term autonomous cleaning robot will be used instead. This could refer to a sweeper, scrubber, or any other applicable type of mobile cleaning machines. This gives the researchers a greater pool of stakeholders to include in this thesis, as even though some of them only are interested in a certain type of cleaning machine, all their viewpoints are of interest.
Figure 1: The sweeping robot that is currently in development at Dyno Robotics.
Limitations
As this thesis mostly goes in depth with the single case study of Dyno Robotics, it is not possible to generalize some of the conclusions in this thesis. For example, the results of the mapping work presented in this thesis should only be viewed as true in its current context. It is hard to know if the cleaning service maps created here holds true in other contexts.
Furthermore, even though implementation is an important part of service design, it is not within the scope of this thesis to implement the design suggestions.
3
Disposition
Chapter 2 – Theory: explains the theoretical framework this thesis is grounded in. It explains
central terms such as design of a service, value and value co-creation, prototyping and robots in services.
Chapter 3 – Method: describes the methodology for the thesis. The chapter describes the
different phases that were included in the thesis and the methods used in each phase.
Chapter 4 – Result: presents the result generated through the research done throughout the
thesis.
Chapter 5 – Discussion: discusses the results within the context of the theoretical framework. Chapter 6 – Conclusion: presents the conclusions of the thesis and attempts to answer the
4
Theory
This chapter lays out the theoretical foundation used in this thesis.
Design of a service
Most people will probably get a clear picture of what one refers to when one says that they design products. While the layman probably mostly thinks of the aesthetics of a product when thinking of product design, and the designer has a more nuanced mental picture that includes more steps of the design process, they can probably easily find a common ground of understanding around the mental picture of a physical product.
However, when one designs a service, this understanding might be more difficult. Where a product often consists of one or more tangible objects, with some digital interactions and services attached to it, a service if often more complex, being mediated by many different people and technologies, working in tandem to deliver messages, products, and performances to the right place at the right time (Holmlid & Evenson, 2008).
In the context of economics, a service can be defined as a value-bringing, economic activity that is intangible, is not stored and does not result in ownership (Investor Words, 2020; Business Dictionary, 2020). As such it is juxtaposed to goods, although it is sometimes referred to as a product since it is offered by a company/provider (Lawrence, Schneider, Stickdorn, & Hormess, 2018). However, there has been a shift in how service research conceptualizes a service. Following a service-dominant logic, service can be defined as a basis of all economic exchange where value is co-created among many stakeholders (Vargo and Lusch, 2016). This indicates that service is not a product extension but a core value co-creating activity in the networks of customers, providers and other stakeholders (Edvardsson, Tronvoll and Gruber, 2011). The service perspective thus puts the focus on understanding customer experience in the context of value co-creation among different actors, and service design is seen as a key to address these challenges (Ostrom et al., 2015)
Service design is an area of design that aims to explore and develop systems that provide a service to users (Mattelmäki & Visser, 2011; van Boeijen, Daalhuizen, Zijlstra, & van der Schoor, 2014). Furthermore, the central focus of service design is the user experience; collaborative participation between the user and other stakeholders is often regarded as essential for making sense of the complexities involved in a service system. It integrates and builds on earlier design fields such as experience design, interaction design, product design etc.
However, there is an ongoing discussion on what the object of service design is i.e. what thing one designs when one designs a service. It has been suggested that the solution to this is to recognize that a service is different things depending on which perspective one has (Sabine, 2009; Kimbell & Blomberg, 2016); the object of service design changes depending on the view one takes. Kimbell & Bloomberg (2016) proposes three different perspectives on the object of service design:
• the service encounter, which focuses on the experience people have as they engage in interactions with touchpoints
• the value co-creating system, which focuses on the dynamic exchanges of resources and processes that achieve outcomes for the actors involved
• the socio-material configuration, which focuses on how resources are made available to different actors. It emphasizes the back end of the service – how the actors are acting and asked to act.
In this thesis, all the above perspectives are important, as they are all needed to capture the multi-facetted nature of the service that is the object of this thesis.
5
Because of the lack of visual appearance in services and the number of stakeholders involved, visualizations play an important role in making the object of service design clearer (Diana, Pacenti, & Tassi, 2009). Visualizations can communicate the idea of a service and make it more concrete for the different stakeholders involved (Bitner, Ostrom, & Morgan, 2008; Diana, Pacenti, & Tassi, 2009; Čaić, Holmlid, Mahr, & Odekerken-Schröden, 2019). Since a service and its elements tend to be more complex to describe than a physical product, it is important to use both abstract and realistic visualizations throughout the whole design process (Diana, Pacenti, & Tassi, 2009). Abstract visualizations include maps and flows, while realistic visualizations include images and narratives, and these can give visibility to the experience or the atmosphere around the service.
The process of designing a service is often highly exploratory and adaptive, and as such there are a plethora of suggested ways to break it down (Morelli, 2006; Meittinen, 2009; Lawrence, Schneider, Stickdorn, & Hormess, 2018). Lawrence et al. (2018) suggests that the four core activities of the service design process is research, ideation, prototyping, and implementation. This thesis uses this approach as a template, but due to limitations of scope, the implementation will be omitted. Furthermore, for reasons explained in chapter 3.1, research and ideation take place at the same time.
As mentioned above, a service can be defined as a core activity of economics, in the sense that it is focused on interactive value creation. But what is value, and how is it created? That is the subject of the next chapter.
Value and value co-creation
Value has different definitions depending on the research focus. Since our thesis explores design work in a service context, the definition used here comes from service-dominant logic, which is a theoretical framework for explaining value creation (Vargo, Maglio, & Akaka, 2008). It states that value is what a customer receives when using whatever product, interaction, resource etc. that is offered by the provider. Furthermore, according to this definition, value only exists when what is offered is in use. This would mean, for example, that according to service-dominant logic, the value of a screwdriver is not received upon its purchase but is instead upon its usage (i.e. when screwing screws).
The shorthand way of describing the type of value that is received upon usage is value-in-use (Grönroos & Voima, 2013), and it can refer to the usage as physical, mental or virtual processes, or it could mean the possession of things.
Service-dominant logic contains a set of axioms used as a basis for understanding services (Vargo, Maglio, & Akaka, 2008). They are the following: (1) Service is the fundamental basis of exchange; (2) Value is co-created by multiple actors, always including the beneficiary; (3) All social and economic actors are resource integrators; (4) Value is always uniquely and phenomenologically determined by the beneficiary; (5) Value co-creation is coordinated through actor-generated institutions and institutional arrangements. Of especial note here is axiom no. 4, as it points out that different people can determine the value of something completely differently.
From a service perspective value co-creation is the creation of value-in-use by the costumers, and other involved actors, and is always co-created through interactions, direct or indirect, between the customer and the providers of the service (Vargo, Maglio, & Akaka, 2008; Grönroos & Voima, 2013; Galvagno & Dalli, 2014). Payne et al. (2008) developed a process-based conceptual framework for understanding and improving the co-creation of value for service providers. Their focus is consistent to service-dominant logic, more specifically the sixth foundation premises (Vargo, Lusch, & Akaka, 2010, p.138) that states that “The customer is always a co-creator of value”. The Payne et. al. (2008) framework describes the suppliers’ value-creation process as:
6
• Identifying co-creation opportunities
• Planning, testing and prototyping value co-creation opportunities with costumers • Implementing customer solution and managing customer encounters and developing
metrics to assess whether the enterprise is making appropriate value proposition
What the authors argue is that the mapping of the value-creation process can aid the identification of co-creation opportunities and they highlight the importance of involving the customer in the entire process of developing products or services. Therefore, the mapping process is especially important for both the customers and suppliers, and interactions between the two since it can help to identify important parameters (e.g. failure points) and improve service enhancement.
The mapping process is connected to another important consideration for value co-creation and that is the role of knowledge and customer and organizational learning (Payne, Storbacka, & Frow, 2008; Vargo, Lusch, & Akaka, 2010; Galvagno & Dalli, 2014). Knowledge can lead to a competitive advantage (Payne, Storbacka, & Frow, 2008) and can be a key to understanding the service (Vargo, Lusch, & Akaka, 2010). Thus, in order to create a product or a service that can co-create value for those involved, the developer needs insight about costumers, and other actors.
The importance of a participatory approach for creativity enabled by service design is discussed by several authors. Mattelmäki & Visser (2011) argue that co-creation focuses on the users and stakeholders and their collective creativity and that creation is a part of co-design, e.g. processes with a collaboration mindset. Mercel, Crul, & Jotte (2016) have done similar work and conclude that co-creation is a process where the firm and the end consumer interact and shares knowledge between themselves. Finally, Sanders & Stappers (2018) mention that co-creation is creativity that is shared between at least two people emphasizing again the importance of collaborative aspect that service design enables. Which activities that are included in co-creation varies depending on what output the researchers wants to receive. Mattelmäki & Visser (2011) and Steen, Manschot, & De Koning (2011) uses different workshops in their cases where their gathered people of interested to work together to solve a given task.
The concept of value co-creation emphasizes the importance and benefits of customer involvement in the service design process and the importance of the participatory approach to spur collective creativity. In order to understand how to gain valuable insights and build knowledge with different actors while designing new services, in the next section we will discuss prototyping as one of the core parts in the service design process.
Prototyping
Prototyping is one of the core parts of the service design process. It can be seen as both a customer and an organizational learning process that serves to evaluate customer understanding of the service purpose and to observe how they interact with it (Payne, Storbacka, & Frow, 2008). According to Blomkvist (2014) there is a distinct difference in service design between prototypes and prototyping; a prototype is referred to as “Any shared physical manifestation externalizing an otherwise internal or unavailable vision of a future situation”. This definition is consistent with the general definition of a prototype (UXL Encyclopedia of Science, 2020), with the caveat that a service design prototype explicitly refers to a situation instead of a physical item. Prototyping, on the other hand, according to Blomkvist (2014), refers to “The use of prototypes to explore, evaluate and communicate ideas in design”. Note that this definition means that prototyping does not mean to construct a prototype, but instead to use an existing prototype (probably constructed or devised by the designer at an earlier point) to explore, evaluate and communicate. A service design prototype
7
can take many different forms, and it is up to the designer to construct or perform one that can generate feedback in the desired format (Servicedesigntools.org, 2020).
Prototype fidelity is an important aspect to consider when prototyping (Buchenau & Fulton Suri, 2000). The closer to the imagined end product/service the prototype is, the higher the quality and accuracy of the feedback can be (Lawrence, Schneider, Stickdorn, & Hormess, 2018) Higher fidelity prototypes have drawbacks though; they naturally require a larger investment to produce, and there is a risk that the inhibit creative discussions, as they can give the impression that certain design aspects already decided, because of the relatively high level of detail. As such, it is important to be deliberate in the choice of fidelity of the prototype.
The above chapters have in broad terms laid out the theoretical foundation for discussing subjects regarding service design. The next chapter will couple this with the fundamental object of this thesis: robots in services.
Service robots
This thesis is focused on an autonomous cleaning robot and the service surrounding it. Therefore, this section will give an overview of the definition and application of service robots as well as implications and what you could gain when implementing robotics in a service context.
2.4.1 Definition and application
According to The International Organization for Standardization a service robot is defined as a robot “that performs useful tasks for humans or equipment excluding industrial automation application” (ISO 8373:2012). The International Organization for Standardization specifies that a degree of autonomy is required in robots and according to them that is the “ability to perform intended tasks based on current state and sensing, without human intervention” (ISO 8373:2012). According to International Federation of Robotics the range of autonomy for service robots is from partial autonomy to full autonomy. Partial autonomy refers to a degree of human-robot interaction, and full autonomy is without any human intervention (Service Robots, u.d.). Further, the International Federation of Robotics categorized service robots further in groups of application, personal or professional.
Depending on the emotional-social and cognitive-analytical range of the task at hand, Figure 2, and the task volume and physical task functionality of the task, Figure 3, human, robot or human-robot practitioners of the task are more suitable according to Wirtz et. al. (2018). If the task is cognitively complex with a high need of emotional support, it is to be expected that the task is delivered by humans with the support of robots (Lariviére, o.a., 2017; Wirtz, o.a., 2018). Further, Wirtz et al. (2018) means that task that are heavy, difficult or even dangerous will be performed by employees that works in teams with robots and therefore the robot can be seen as an autonomous and smart tool that aids the employees in, for them, unpleasant tasks.
According to Park & Pobil (2013), there are several aspects that can affect a user’s intention to use a service robot. First one is the ease of use. The authors meant that if the service robot is perceived as easy to use, a user is more likely to use the robot more than the first time. Another factor is the perceived usefulness of the service robot. A user intention to use a robot could decline if the robot does not feel useful in its tasks (Park & Pobil, 2013).
This section has brought forth the definition of robots with the intended use in services. The section has also discussed in which situation a certain level of automation in robots is more suitable. This information is intended to be used to explore the autonomous cleaning robot’s position in a cleaning service system. The next section talks about implications of introducing robots in a broader spectrum.
8
Figure 2: Service delivery based on the complexity of emotional and cognitive tasks (Wirtz, o.a., 2018).
Figure 3- Service delivery based on volume and heterogeneity of physical tasks (Wirtz, o.a., 2018). 2.4.2 Introducing service robots
Cleaning robots, especially vacuum cleaning robots have a long and widely-used history in the market according to Prassler et al. (2000). There are many gains for implementing service robots, or robotics in general (Bishop, 2006; Curcio, 2012; Wirtz, o.a., 2018; Huang & Rust, 2018). If cleaning robots worked collaboratively with human cleaners, the robot could take over repetitive tasks that exist in the cleaning industry. Then the employees can focus on tasks that robots cannot perform. (Curcio, 2012) A cleaning robot can also be programmed to identify the most effective way of several scenarios to clean a surface and can behave identically through the tasks (Bishop, 2006; Curcio, 2012). Curcio (2012) also mentioned that cleaning robots could operate in conditions that are not suited for human employees, e.g. strong wind. Robots also lack the ability to make human errors and the ability to feel tiredness (Huang & Rust, 2018).
Looking on the system level, robots could reduce the costs for the service provider. The highest costs related to the robot rests in the developing- and manufacturing phase (Wirtz, o.a., 2018). Some additional costs could occur for physical robots with time, but Wirtz et al. argues that it is a fraction of the costs for increasing the size of the staff. They argue that scaling up a service staff, especially high-performance staff for a better service experience for the costumer, has high costs.
According to Gumbel, Chui, & Lund (2018) “Automation will displace many jobs over the next 10 to 15 years.” But the authors also argued that just because jobs get more and more automated does not mean that there will be less jobs for humans to do. They imply that the tasks will shift, and the employees will have other working responsibilities. The authors also
9
point out that history has shown this and that technology has created new jobs in the past, example when the ATM:s were introduced. Instead of taking jobs from the human employees, they created more. The reason was that by implementing the technology, the bank business became cheaper to manage, and could be expanded throughout the country, and thus create more job opportunities. The challenge, however, is that a lot of people will lose their jobs, and somehow they need to gain new skillsets that are suited for other jobs. (Gumbel, Chui, & Lund, 2018)
This section has talked about the implication of implementing service robots in a service context. This will be used as a foundation in this thesis to find more aspects for implementing robotics in a cleaning service. The next section describes the different levels of autonomy when it comes to service robots.
2.4.3 Levels of autonomy in service robots
There are many definitions of autonomy, but what is common in many definitions is the capability of adapting to surrounding environment and be able to reach a goal without assistance (Beer, 2014). But a robot or a machine is seldomly either autonomous or non-autonomous, instead it seems better to define its autonomy along a scale of different levels of autonomy (Beer, 2014). One such scale that is widely used today is the The National Highway Traffic Safety Administrations (NHTSA) six levels vehicle autonomy (Shuttleworth, 2019) were, for example, level 0 indicates no automation, level 1 indicates automation of a single function, such as automatic transmission, and level 5 indicates full autonomy in all situations. Beer (2014) suggests a Taxonomy of Levels of Robot Automation in HRI (LORA), which defines different levels of automation of service robots. It contains 10 levels of relatively fine granularity, i.e. level 6 is “shared control with human initiative”, and level 7 is “shared control with robot initiative”. The highest level is level 10, where the robot is “fully autonomous”.
10
Method
This chapters describes the different methods used in this thesis. First, an overview is given that briefly presents which phases were included and why the approach was selected.
The process overview
This thesis explores what different stakeholder’s value in a current and a future cleaning service. As mentioned in chapter 2.1, Design of a service, this thesis assumes the idea that a service can beneficially be viewed from different angles, allowing one to be more specific when e.g. discussing what is valuable about a service. As such, the answer to what is valued or what hinders a future cleaning service changes depending on which perspective one views it from. The different prototyping techniques used throughout this thesis deliberately probed the current and future cleaning service using the different perspectives of the value co-creating system, the socio-material configuration, and the service encounter (Kimbell & Blomberg, 2016). Table 1 below clarifies how each perspective relates to the service and which prototyping techniques explored which perspectives.
Table 1 – Which aspects of the service that the different perspectives of the value co-creating system, the socio-material
configuration, and the service encounter relates to and which prototyping techniques were used to explore them.
The overall process of performing this thesis was divided into three different phases: Mapping & Ideation, Conceptualization and Prototyping, which can be seen in more detail in Table 2. Most of the activities were co-creative together with stakeholders. In the Participants column the number of workshops can be seen, including the number of participants at each workshop. For example, the Mapping & Ideation workshop in total had 4 workshops, with the number of participant N=1,2,3… varied. Note that even though Affinity diagram is a data analysis technique and not a prototyping technique, it is shown here for the sake of demonstrating the process of the thesis. The first activity, however, was a short pre-study, consisting of a co-creative workshop together with Dyno Robotics, aiming at generating a preliminary list of relevant actors and stakeholders that could be involved in the future cleaning service system. Those stakeholders were: manufacturer of cleaning machines, property owner, property users, retailer of cleaning machines, cleaning company, end-user of robot, garbage collectors and policy makers.
11
Table 2: The different phases of this thesis. Process stage Prototyping
technique
Purpose Format Participants
Mapping & Ideation Contextual stakeholder value mapping Function ranking Mapping the current and future service activities Value discovery and ideate future state of service Co-creative workshops Property Owner 1 (N=1) Retailer (N=1) Outdoor Cleaning Company (N=3) Property Owner 2 (N=1) Conceptualization (Affinity diagram) Generate concepts based on the data and insights gathered during the Mapping & Ideation phase
Workshop Researchers of this thesis (N=2) Prototyping Desktop walkthrough Evaluate the concepts that was generated in the Conceptualizat ion phase and identify potential problems on a system level Co-creative workshops Property Owner 3 (N=2) Property Owner 2 (N=2) Experience prototyping Test and evaluate the user experience on an interaction level Individual prototype testing sessions Property Owner 2 (N=3) Outside Cleaning Company (N=6) Property users (N=8) Service Advertisement Testing the desirability and perceived value of a service offering Questionnair e sent by email to key stakeholders Property owner 1 (N=1) Property owner 2 (N=2) Property owner 3 (N=1) Retailer (N=1) Outside cleaning company (N=1)
12 Policy maker (N=1) Synthesis of gathered data (Affinity diagram) Analyze the gathered insights and identify common themes
Workshop Researchers of this thesis (N=2)
In this thesis, the typical design activities of research and ideation take place during the same workshops and are instead called Mapping & Ideation. Due to Dyno Robotics already having a physical prototype, and that the project from the start already had a rough sense of direction, the researchers had the prerequisite material and knowledge to be able to conduct activities related to both the phases in the same workshop. Furthermore, the researchers could not demand that the stakeholders set aside time for two separate occasions.
All phases had a collaborative activity bound to them, except for the Conceptualization phase. Instead of having collaborative activities, the aim of the Conceptualization phase was to analyze the data collected and synthesized in the Mapping & Ideation phase, to in order generate concepts which could be prototyped in the prototyping phase.
Data gathering and insight generation
At its core, the purpose of each prototyping method used has been to gather data. Data has been gathered during each meeting in the form of written notes, audio recordings and pictures, and has then been analyzed through means of discussion, and turned into insights. They are distinct pieces of data with a short text added in order to contextualize them. Each prototyping session has generated circa 4-8 insights, which have been written into their respective insight documents, see Appendix 1 for all the insight documents gathered in this thesis. An insight can for example be that “retailers value product flexibility”, with the context that product flexibility gives them more wiggle room when selling a product, since they easier can find a balance between customer budget and needs.
In some cases, when a prototyping session has generated exceptional amounts of data, affinity diagrams (American Society for Quality, 2020; Interaction Design Foundation, 2020) have been used to cluster the data into manageable groups of common themes, which then have been converted into insights. In either case, insights have been the end product of each prototyping session. Table 3 below shows the process of insight generation.
Table 3: The process of insight generation Data recorded
during meeting Method of analysis after meeting Insight document
Written notes Discussion
Audio recordings Affinity diagram
13
The different methods used to gather the data shown in the following chapters.
Mapping & Ideation
The core activity of the Mapping & Ideation phase was a co-creative workshop that was held with different stakeholders. The stakeholders involved in this phase can be seen in Table 2. The workshop generally took 1-2 hours to conduct and was divided into three activities:
current state mapping activity, future state mapping activity and an ideation activity. All
these activities also served as a medium in which to perform semi-structured interviews to gather qualitative data about the stakeholders view on the current and a possible future cleaning service.
Since each stakeholder was unique, the specific questions prepared different somewhat between each workshop. The stakeholders involved in the workshops had been chosen based on their direct involvement and influence of the cleaning service system.
The researchers assumed that the respective data about the cleaning service system collected from each stakeholder would probably be biased around each respective stakeholder. Thus, further analysis and synthesis of data would be required after all of the workshops had been held, to encapsulate the whole service system.
3.3.1 The workshop
Here the different activities of the workshop are explained. The current state mapping activity
The goal of the current state mapping activity was to gather knowledge about the current cleaning service. The method used was an adapted version of the contextual value network mapping activity (Čaić, Holmlid, Mahr, & Odekerken-Schröden, 2019) and was presented like this:
Using the list of stakeholders generated from the pre-study, actor cards were prepared to symbolize stakeholders and actors that the authors deemed reasonable to be involved in the cleaning service system. see Figure 4 for an example. Several actor cards were printed without a title to allow for custom cards to be created during the activity by writing on them.
14
During the mapping activity the participants were asked to map out the cleaning service system using the actor cards. To help them start, they were asked to imagine which actors a cleaning machine travels through, and to place out the relevant actors and elaborate on their relationships. Below, in Figure 5, an example of a mapping of the cleaning service system in progress using the actor cards can be seen.
Figure 5: An example of the current state mapping activity in progress. Icons made by Freepik from www.flaticon.com
During this activity, the focus for the participants was to answer the following questions: • What actors are involved in the cleaning service?
• How do the actors relate to each other in the cleaning service? The future state mapping activity
After the current state mapping activity had been performed, the next task was for the participant to place the robot into the current state map and correct any inconsistencies that appeared, thus creating the future state map. See Figure 6 below for an example of the future state mapping activity in progress.
Figure 6: An example of the future state mapping activity in progress. Notice that in this particular example, the only difference between the future and current state service map is the addition of the robot actor in the upper middle section of
the picture. Icons made by Freepik from www.flaticon.com
During this activity, the focus for the participants was to answer the following questions: • Where in the current network map would the autonomous robot fit?
• How would the autonomous robot be used?
15
The ideation activity
In preparation for this activity, a list of possible functions that the robot could possess was created together with Dyno Robotics. Cards were then created that visualized these functions, like the cards used in the mapping activities. See Figure 7 below for an example. The goal was to, by having the participants grade the functions while simultaneously discussing them, find which functions they value. The facilitators introduced and explained the cards to the participants, and they were instructed to arrange them according to importance. Supplementing questions were asked to get a deeper understanding on why certain functions were perceived more important.
Figure 7: Examples of function cards. Icons made by Freepik from www.flaticon.com
During this activity, the focus for the participants was to answer the following questions: • Which function should the autonomous robot have and why?
• Which functions are less important for the autonomous robot and why? 3.3.2 Map generation and analysis of the function cards activity
After each workshop, the collected data (notes, photos and audio recordings) was analyzed through discussion and synthesized into a current and future state partial service map, resulting in one set of current state and one set of future state service map per stakeholder. An example of a current and future service map, as seen from the perspective of one stakeholder, can be seen below in Figure 8 and Figure 9. The full set of current and future service maps, as seen from the perspectives of the different stakeholders, can be found in Appendix 2.
The function card activity was also analyzed and put into a combined table with comments from different stakeholders, see Appendix 3.
16
Figure 8: An example of the cleaning service as seen from one stakeholder
Figure 9: An example of the future cleaning service as seen from the perspective of one stakeholder
After all the workshops had been performed and the researchers had created system maps for each stakeholder involved in the Mapping & Ideation phase, a complete current state service map was created from the partial service maps. This was achieved by first drawing an outline of the complete service system, and then filling it in with the details found in each partial service system map.
17
Conceptualization
The conceptualization phase was relatively short and mainly consisted of a workshop that only the researchers took part in. The aim of the workshop was to summarize what had been learned in the Mapping & Ideation phase about what different stakeholders value in a cleaning service and based on this create concepts to test in the Prototyping phase.
An affinity diagram (American Society for Quality, 2020; Interaction Design Foundation, 2020) was used to cluster the insights that had been gathered. The insights that had been continuously gathered were written down separately on post-it notes. They were then clustered into groups with other insights that seemed to share a common theme. Figure 10 below shows a snippet of the Affinity diagram that was created.
The insight clusters were then used as a base from which to discuss and generate concepts. Many different concepts with minor variations were discussed, in the end however, only two concepts were of interest – The Collaboration Robot and The Nighttime Robot.
Figure 10: An example of a part of an Affinity diagram.
3.4.1 The Collaboration Robot
The main idea with The Collaboration Robot concept was that a worker and the robot work alongside each other. While the robot was cleaning the streets, the worker can take care of green areas. They are not required to work in immediate proximity of each other, but if the robot had any issues, it would notify the worker through his or her smartphone. This concept was partly derived from the fact that some stakeholders expressed a worry of vandalism or theft if the robot worked alone.
This concept would require a well thought out user experience; it is important that it does not negatively affect the work performance of the worker that collaborates with it.
18
3.4.2 The Nighttime Robot
One of the most lucrative aspects of an autonomous cleaning robot, according to the stakeholders interviewed, was the fact that it could perform work without the involvement of any individuals. One of the ultimate embodiments of this would a robot that worked throughout the night, both without the help of people and without being in the way of individuals.
This, however, placed rather high requirements on the robustness of the systems involved. For example, the robot would perhaps need to be able to communicate with door opening mechanisms and be ignored by alarm-triggering motion detectors at the area of operation.
Prototyping
Prototyping was used in this project both to evaluate concepts and ideate together with the participants. This was an iterative process, and the plan was to evaluate concepts of increasingly fine granularity. There were three parts of the prototyping phase: Desktop walkthrough, Experience prototype and Service advertisement.
3.5.1 Desktop walkthrough
Desktop walkthrough is a service design technique used to collaboratively build a miniature environment to construct knowledge about a specific service (Blomkvist, Fjuk, Annita, & Vasilisa, 2016). In more practical terms, with the use of e.g. Lego, one collaboratively (e.g. together with one or more stakeholders) constructs and plays out an aspect of a service, such as a service encounter or the back-end system supporting a service encounter. This prototyping method is great to do early in the process, as it requires relatively little investment or prior knowledge to perform (This is service design doing A, 2020).
In this thesis, Desktop walkthrough was used to evaluate the two different concepts that resulted from the conceptualization phase: The Collaboration Robot and the Nighttime Robot. It was also used as a way to get insight on how the different stakeholders (Table 2) would use the autonomous cleaning robot and to identify potential problems in the future service. The stakeholders participating in the workshop were chosen for their knowledge about the use of cleaning machines and about what problems could arise.
Lego was used to represents different aspects of the service, e.g. the robot, vehicles, and people. Maps of the area where the stakeholders operate were printed out and used as the location of the service. Figure 11 illustrates the setup for the session and Figure 12 shows an example of a desktop walkthrough in action. Facilitator 1 introduced the participants of the tasks, while facilitator 2 documented the session through written notes.
The process of the sessions was as followed:
1. The participants got an introduction from the facilitators about the autonomous cleaning robot
2. The props were introduced and explained
3. The participants were instructed to demonstrate with the props how the robot would be used
4. Different obstacles that could occur were introduced by the facilitator and the participants were urged to find solutions to them
19
Figure 11: Setup for the desktop walkthrough.
Figure 12: Example of a desktop walkthrough in action.
3.5.2 Experience prototype
An experience prototype is any kind of representation, in any kind of medium, that is designed to understand, explore or communicate what it might be like to engage with the product, space or system that is being designed (Buchenau & Fulton Suri, 2000). The idea is to let designers and clients have informative and personal experiences in order for them to feel greater empathy with both the people who will be affected by their decisions, and the experiences users may face. The above definition of experience prototyping is quite broad. As such, lots of different types of prototypes could be defined as an experience prototype – as
20
long as the goal is to capture the experience of the user. In this thesis, the experience prototype was in the form a usability test of an interactive prototype of the physical robot.
Usability testing involves measuring the efficiency of a prototype in regard to task completion (John Wiley & Sons, 2002). There is a myriad of ways to measure this efficiency e.g. time-to-completion, number of mistakes etc. Besides the quantitative efficiency of the prototype, one can also evaluate the more qualitative aspects; there are standardized questionnaires, such as SUS, that can be used to evaluate the level of user satisfaction.
In the experience prototype that was carried out during this thesis, participants got to interact with the cleaning robot using three different modes of interaction: voice, a display mounted on the robot, and via a smartphone. The reason why the experience prototype contained these specific modes of interaction, as opposed any other modes, was because they were the ones that had been mentioned and discussed with stakeholders in the workshops before. Furthermore, they also fit well with the prototype that Dyno Robotics provided.
The test was performed to find valuable aspects for each mode and what mode gives the most valuable experience to a future user of the autonomous cleaning robot. The process of the three modes of interaction is described in more detail further down.
The participants of the experience prototype were selected for their future interaction with the robot, see Table 2. for the total number of participants. The motivation for including cleaning workers in the experience prototype was that they could be seen as the future users of the robot, and therefore the users who would interact with the robot the most. A potential future scenario could be for the robot to operate in public places and would then come across the people that live and work in the property. Therefore, the facilitator of the thesis thought that there were valuable insights to be gathered from the property users as well.
The modes of interaction were tested using three different scenarios. The three scenarios were repeated for every mode of interaction. The scenarios were created together with Dyno Robotics during a small co-creative workshop that included brainstorming on the theme of possible things that either the robot or the user could want to communicate to the other party. The collection of these ideas were then clustered using the Affinity diagram technique (American Society for Quality, 2020; Interaction Design Foundation, 2020), and from that, the three different scenarios were specified. The three scenarios explored real-time instances that could occur between a future user and the autonomous cleaning robot:
• Scenario 1: Schedule of a cleaning program for a later time
• Scenario 2: The robot notifies the participant that it needs help with garbage disposal • Scenario 3: The participant sees a dirty area that needs to be cleaned right away
See Appendix 4 for a more in-depth description of the three scenarios. The order of the scenario remained the same for each mode of interaction, but the order of the modes varied for user to user, see Table 4 for an example of this. This to be able to evaluate more reliably. The facilitators assumed that participants learned the system after each mode of interaction, and they wanted to get data on when the participant experienced the system for the first time, for all three modes of interaction. And therefore, a different order for the modes was decided.
The voice mode had one test scenario before the first scenario. Interaction with technology by voice is not that common for most people and therefore the test scenario was used to introduce people to the technology. During the test scenario, users introduced themselves to the robot and answered the question the robot would ask.
The physical prototype of the robot, Figure 13, was constructed by Dyno Robotics and consisted of a modified sweeper machine. The application interface for the display and smartphone was designed using Adobe XD and developed using Unity. The user interface that was used was not the main factor to evaluate during the testing and therefore the design of the interface was not researched in more detail in the thesis work and a more practical user interface was used.
21
Table 4: Example of the order of the modes of interaction for some of the participants. For every mode, scenario 1-3 was performed with the same order.
Participant Order of the modes of interaction
1 1. Voice 2. Display 3. Smartphone 2 1. Display 2. Smartphone 3. Voice 3 1. Smartphone 2. Voice 3. Display
Figure 13: The experience prototype being carried out.
All sessions were video- and audio-recorded and one of the facilitators took written notes during the tests. The participant got approximately 15-30 minutes to conduct the entire test, including the interview and questionnaire-part. The experience prototype was done in three different occasions during a time period of three weeks.
Data analysis
To be able to evaluate the experience prototyping, three types of methods was used, system usability scale (SUS), (Brooke, u.d.), performance success and semi-structured interviews. After every mode of interaction, the user got to fill in a questionnaire (SUS) with ten questions regarding the concept in its entirety. A value between 0-100 was then calculated for every questionnaire that represented the quality of the concept as of today. How the SUS score was interpreted is displayed in Table 5.
22
Table 5: (System Usability Scale (SUS) Plus, u.d.)
SUS Score Letter Grade Adjective Rating
>80,3 A Excellent
69-80,3 B Good
68 C OK
51-67 D Poor
<51 F Awful
The second measurement taken was performance success (time-to-completion). Every scenario done by the participants were timed and documented. The time of the participants were then measured by comparing their execution time with the time from an expert user, a user with previously knowledge about the system. In this case the expert user was one of the facilitators. The experts time was then multiplied with two to create a threshold value for the participants time to pass.
After the prototyping session, each SUS-document was analyzed and put into tables, as well as the execution time for each user. This together with notes taken from the session, as well as the answer to questions asked was translated into insights and insights document with the same approach as explained in Data gathering and insight generation.
3.5.3 Service advertisement
Service advertisement is a prototyping technique that involves creating a prototype advertisement in order to probe the perceived value of a new offering (This is service design doing B, 2020). It is a sort of “fake it before you make it”-approach, used to evaluate the core value proposition of a service. It can take many different forms, such as a flat poster, video or even an interactive web site. Furthermore, when creating a prototype service advertisement, it is important to go beyond simply describing facts about the product/service, and instead try to convey an emotion to the audience.
In this thesis, the Service advertisement was used to test the desirability and perceived value of a cleaning service that uses autonomous cleaning robots. The researchers designed an advertisement poster, Figure 14, that illustrated the service offering. The content on the poster was produced from insights that had been gathered from the previously meetings with stakeholders. The poster was then sent by email to the following key stakeholders that had been included in the thesis work up until that point: retailer, property owner, cleaning company and policy maker. A questionnaire with relevant questions was sent together with the poster for the stakeholders to answer. The purpose of the questionnaire was to get insight about whether the presented service concept was perceived as valuable according to the stakeholders and if that was not the case, what could make the offer more valuable and more desirable.
23
24
Synthesis of gathered data
During the synthesis of gathered data phase, all the insights gathered through the thesis was considered, see Figure 15 to see where the insights originated from. The aim of the phase was to analyze the insights and identify common themes among them.
Figure 15: Insights gathered through different phased and methods.
The phase began with the researchers writing down the insights on individual notes and grouping them in a by discussion and using the affinity diagram, where the insights with common themes were grouped together, such as. What was searches for was common themes among the insights that, for example the insight People found the technology to be interesting and All stakeholders seem to find the idea of an autonomous cleaning machine valuable were grouped together.
After the first clustering, another was performed with the aim of identifying what insight groups could be seen as Aspects of value for the stakeholder, and what could be seen as risks that could hindered the creation of the identified values, and that would need to be considered during the development of an autonomous cleaning robot in a service context.
The result from the phase was: • 8 Aspects of Value
• 7 Design parameters • 4 Design problems
25
Results
In shorthand, this thesis aims to answer what a current cleaning service looks like, what a future cleaning service that uses autonomous robots could look like, what and in what ways different stakeholders could value said service, and how one should approach developing it. This chapter presents the relevant results to answer those questions. It is divided in to three sections:
• Current and future service map, which presents three different maps that describe the current cleaning service, the value exchange network in the current service and the suggested future service that uses autonomous robots
• Aspects of value, which presents the different ways in which stakeholders could value a future service that uses autonomous robots
• Design parameters and design problems, which presents important factors that could either enable of inhibit the ability of the service to deliver its intended value.
Current and future service map
In the first phase of this thesis, several co-creative workshops were conducted to get knowledge and map out how the current cleaning service system looked like, as well as ideate how a future cleaning service system that uses an autonomous cleaning robot could look like. This resulted in two separate service maps as well as a value exchange map that illustrate what values are exchanged between the involved actors and stakeholders. The results are described in more detail below.
26
4.1.1 Current service map
The service map, Figure 16, illustrates what the service system looks like according to the research. While it shows the most important connections between the most important stakeholders, it is worth the note that it is a simplification regarding the number of steps taken in the different processes involved.
The map illustrates how the different actors and stakeholders relate to each other and whether they have a primary, secondary, or tertiary role within the service. The different lines, transport of cleaning machine, transport of waste, cleaning machine maintenance service and property user satisfaction feedback illustrates from whom a certain flow originated and to whom it transfers through. A larger version of this map can be found in Appendix 5.
27
4.1.2 Value exchange map
Figure 17 below shows the value exchange map, which complements the service system map by explicitly showing the different values that different stakeholders and actors give each other. For example, while the retailer receives machines to sell from the manufacturer, what the manufacturer receives in return is a broader clientele.
28
4.1.3 Future service map
Figure 18 below illustrates the suggested future service. Compared to the current cleaning service, this service has the added elements of policy maker, who may regulate autonomous entities, server which connects the robots with the cleaning companies and Dyno Robotics, as well as making Dyno Robotics the manufacturer of the service. Note that this map, just as the current service map, is a simplification. A larger version of this map can be found in Appendix 5.
Figure 18: The future service map.
Aspects of value
As mentioned above, one of the aims in this thesis is to answer in what different ways different stakeholders could value a future cleaning service that uses autonomous cleaning robots. The phase synthesis of gathered data has shown that the value of a future cleaning service that uses autonomous cleaning robots can be divided into the following different aspects:
• Cleaning work, referring to the perceived value of the cleaning work that the robot performs.
• Relief of repetitive strain injuries, referring to the perceived value of the reduction of repetitive strain injuries that the robot could bring.
• High-tech branding, referring to the perceived value of the possible high-tech branding that the robot could bring.
• Technological enthusiasm, referring to the perceived value of the enthusiasm using and owning new, or otherwise desirable technology, could bring.
• Cleaning consistency, referring to referring to the perceived value of the result of the cleaning work being consistent, as opposed to it varying.
