S.M. Media Arts and Science
Massachusetts Institute of Technology, 2007
B.F.A. Industrial Design
Rhode Island School of Design, 1995
Submitted to the Program in Media Arts and Sciences, School of Architecture and Planning
in partial fulfillment of the requirements for the degree of
Doctor of Philosophy at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY
© Massachusetts Institute of Technology 2013. All rights reserved.
September 23,, 2013
Program in Media Arts and Sciences
Professor of Media Arts and Sciences
MIT Program in Media Arts and Sciences
LEGO Papert Professor of Learning Research
Academic Head, Program in Media Arts and Sciences
Submitted to the Program in Media Arts and Sciences,
School of Architecture and Planning, on September 23, 2013
in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy in Media Arts and Sciences at the
Massachusetts Institute of Technology
Sartorial Robotics is a method of merging fashion theory and robotics through the design and development of robotic systems. These systems facilitate interaction and play as well as mimic the materiality, aesthetics, and construction techniques of textiles, apparel and fashion. This will enhance the social aspects of human-robotic interaction and assist in how we situate robotics in our lives and cultures.
Building upon a history of robot aesthetics and a formulaic approach to analyze and understand fashion, a series of design principles for Sartorial Robotics were established and applied in the research and development of robotic systems that utilize the human-centric system of clothing to create robotics for human-robot social interaction. The Group Identity Surface is a soft-architecture system utilizing thermochromic textiles and computer vision to facilitate human-machine teammate building. Zipperbot, a robotic continuous closure for fabric edge joining, was developed to explore autonomous control of a sartorial gesture and performed as a wearable robot which was evaluated through social interactions.
Clothing is a uniquely human pursuit and is nearly universal in its adoption and use. It plays a prominent role in our individual cultures transmitting a mixture of social signals and meanings through the semiotics of fashion. It is through this performance of assemblage of fabric surfaces we reconfigure ourselves and our identities. Merging robotics and fashion within the practice of Sartorial Robotics will enhance the explorations of identities for both humans and robots.
Professor of Media Arts and Sciences
Massachusetts Institute of Technology
Associate Professor of Media Arts and Sciences
Massachusetts Institute of Technology
Ute Meta Bauer
Associate Professor in Art, Culture and Technology
MIT Department of Architecture
I would like to express the deepest appreciation to my advisor Professor Cynthia Breazeal for her constant support, her encouragement and vision which provided the opportunity for me to pursue this unique combination of concepts and ideas. I am incredibly thankful to my dissertation committee, Leah Buechley and Ute Meta Bauer whose enthusiasm, creativity and encouragement provided the connections to combine all these diverse concepts into a cohesive whole. Thank you.
I am grateful to the entire MIT community specifically the Media Lab and to my longtime advocates, Krzysztof Wodiczko and my first advisor Chris Csikszentmihàlyi who supported my work at every step along the way. I wish to thank the MIT faculty specifically Hugh Herr, Hiroshi Ishii, Joe Paradiso, Ted Postal and Dava Newman. I would also like to thank friends and colleagues, Dr. Margot Krasojevic, Sung Ho Kim, James Cain, Seth Riskin, Michael Enos and Jim Gouldstone. I am particularly appreciative to the Personal Robots Group for all their support, assistance and passion: Angela Chang, Nick de Palma, Adam Setapen, Jin Joo Lee, Philipp Robbel, Peter Schmitt, Siggi Örn Aðalgeirsson, David Robert, Natalie Freed, Kenton Williams and Kristopher Dos Santos. A special thank you to Polly Guggenheim who keeps the group running and helps us all traverse this epic process. I wish to thank my friends and colleagues from Comp Cult: Kelly Dobson, Noah Vawter, Nadya Peek, Annina Rüst, Gemma Shusterman, Mako Hill, Josh Levinger, Ryan O’Toole, Alyssa Wright, Ayah Bdeir, Sara Ann Wylie and Stephanie Gayle. I would also like to thank John DiFrancesco and Tom Lutz and a particular thanks to Linda Peterson for keeping me on track.
Most of all I would like to thank my family, it was only with them that all of this was possible. I cannot thank enough, my father, John Whiton and mother Marcia Whiton, they have imparted upon me their wisdom and commitment to the pursuit of knowledge which has guided me not only in this work but in life. I wish to thank my sister, Kristina Whiton and her husband Tom O’Brien who, through the whole process kept their faith in me. I am especially thankful to my wife, Yolita Nugent who has not only offered her unwavering support but has been my constant collaborator and inspiration through all of this work. I am forever grateful to my son Reece, his boundless curiosity and creativity inspires me every day to maintain a sense of wonder in the world.
Table of Contents
Chapter 1 Introduction 13
Surfaces and Playing with Identity 15
Shifting Methodologies for Robotics 17
Thesis Structure 20
Chapter 2 Background and Related Work 21
Toward Development of Surface Aesthetics in Robotics 21
Electronic-textiles and Fibers that Sense, Actuate and Compute 31
Soft-architecture Robotics 33
Robotics and Clothing 35
Chapter 3 Wearables for Robots Prior Work by Author 41
Chapter 4 Evaluation of Sartorial Cues and Robotics 49
Experiment Design and Procedures 52
Gender Neutral Control Biased Feminine 54
Masculine Sartorial Cue rated higher Masculine Traits and lower on Feminine Traits 55
Feminine Sartorial Cue rated higher Feminine Traits and lower on Masculine Traits 55
Comparing Masculine and Feminine Traits 56
Individual Masculine and Feminine Trait Shifts 57
Chapter 5 Design Principles and Framework 60
Public and Personal Sartorial Signals 62
The Formulaic Approach to Analyzing Fashion 63
Core Functionalities: Sensing, Actuation and Computation 65
Sensing: Sartorial Detection 65
Sartorial Actuation 66
Sartorial Computation 66
Utilizing Apparel Techniques 68
Diversity of Robot Forms 68
Telepresence Robots 72
Sartorial Robotics on People 72
Develop Series of Demonstration Projects for Sartorial Robotics 73
Chapter 6 Group Identity Surface 74
Group Identity Surface 75
Electrically Controlled Thermochromic Narrow Fabric 76
OpenCV Programming 79
Interface PCB Hardware 80
Chapter 7 Zipperbot 83
Zipperbot: A Robotic Continuous Closure for Fabric Edge Joining 84
Mechanical Design 85
The Motor 88
Position Detection 89
Printed Circuit Board Design 91
IOIO Board and Android 93
Zipperbot in a Hobble Dress 93
Evaluation of Autonomous Robotic Sartorial Gestures using Zipperbot 95
Research Questions 96
Materials and Methods 96
Experiment Design and Procedures 96
Positives and Negatives in a Broad Range of Perceptions 99
Perceived Body Posture 100
Questions of Control 101
Analysis of Body Language and Verbal Utterances 102
Chapter 8 Future Work 107
Agency and Control 107
Social Performance 108
Sartorial State versus Sartorial Gesture 109
Designing with Gestures 110
Multiple Zipperbots 111
Chapter 9 Contributions 112
Soft-architecture Robotics 112
Interdisciplinary Dialogues 113
Sartorial Nonverbal Communication 113
Table of Figures
Why in the pursuit of robotics, or more specifically social robotics, should we make an attempt to integrate fashion theory? This question requires a multifaceted explanation. At first glance the two individual disciplines seem miles apart. However, as we progress through technological changes that shift computation into electronic-textiles and fiber-electronic devices, and the field of robotics transitions from industrial manufacturing and laboratory research into the consumer market and thus into our homes and our personal and social lives, the research space for robotics must expand and these two disparate fields of fashion and robotics are presented an opportunity to merge.
QB Avatar is a telepresence robot which is meant to give people a virtual presence anywhere they physically cannot be. As their promotional material advertises, it “enables you to be where your heart needs to be, but you can’t.” It is therefore intended as a social robot, a robot that interacts and communicates with people and is thus expected to demonstrate social behaviors. At a wedding where an elderly relative was unable to physically attend, the QB Avatar robot was employed[Emi12]. However, this real-life, highly social application presented an issue. The attendee would be able to virtually attend but would not able to take part fully in the pre-wedding costume party. The robot could not dress up. Additionally, they would not be able to partake in the formal dress customary of the wedding day. In sociology, it is understood that our corporeal bodies perform a critical role in our social interactions. Graham Scambler, health theorist and medical sociologist, in discussing the body during debilitating illness and health-related stigmas states:
to be acknowledged as competent social performers we have to be able to give the impression of some degree of control, use and presentation of our bodies.
Our bodies fail us not only in illness, but consider the social ramifications when we uncontrollably blush or inadvertently stutter. This issue of bodily control actually extends beyond our bodies to our surfaces, specifically our clothing; wardrobe malfunctions and other fashion faux pas by which we are socially judged, like a wrinkled business suit or stained dress, will have significant social consequences, i.e. lack of competent social performance.
The solution for the virtual attendee was a photographic cardboard cutout attached to the robot1. The lack of customizable or even adaptable surfaces on the QB Avatar robot spurred some very creative ingenuity which demonstrated the need and desire to play with surface for social expression and interaction Figure 1-1.
Figure 1-1. QB Avatar telepresence robot modified with photographic cutout to represent separate outfits of the virtual attendee of a wedding.
Despite all the technological advances of the telepresence robot it lacked the basic mechanics of clothing and fashion which we effortlessly choreograph every day when we dress.
Clothing performs a prominent role in our social presentation and transmits a variety of social signals and meanings through the semiotics of fashion in a form of sartorial signaling. Roland Barthes suggests fashion is play, with one of the most profound of human questions “Who am I?”[Bar90]. It is through this performance of assemblage of fabric surfaces, play, we reconfigure ourselves and our identities. My hypothesis is that a robotic surface, which facilitates interaction and play as well as mimics the materiality, aesthetics, and construction techniques of existing textiles and apparel, will enhance the social aspects of human-robotic interaction and assist in how we situate robotics in our lives and cultures.
Surfaces and Playing with Identity
The morphology of clothing is in many ways a modular manipulation of surfaces and it is through this manipulation that we are able to explore and express identity of self, culture and social structures. However, even without clothing, human beings have expressed their identities through their very own skin. Researchers and roboticists in their pursuits of robotic skin generally mean coverings or surfaces designed to sense or detect mechanical pressure or a simple flexible covering. However, skin as we know it in living organisms and of course human skin is a more complex multilayered living organ. Functionally, it acts as a layer of protection from the environment and helps maintain a constant body temperature with the production of sweat. Our skin is responsible for the metabolism of vitamin D when exposed to ultraviolet medium wave light. It has tiny hair follicles, nerve fibers, arteries, and veins, and through capillary dilation can indicate an emotional response, we blush. The melanocytes stratum in the epidermis layer of our skin provides pigment and color, something that can have significant socio-cultural impact. Culturally our skin is pierced, cut and scarred, tattooed and wrinkled from our passage through time[Ahm01]. Through an analysis of human skin beyond visual and sensory experiences we can observe that our skin is complex; our skin is alive and is not simply a static surface; it is a system, a system engaged with meaning. From the perspective of the materiality of the human body, our skin is not a simple covering or surface boundary; it is a functioning part of our whole and an interface to our identities.
Our clothing is an extension of our skin, both in its functionality and as an expression of identity. However it offers us quite a bit more in its flexibility and modularity. We can don and doff it, change or alter it, or discard it entirely with relative ease particularly in comparison to our skin. We can also employ it to experiment with other objects.
Figure 1-2. A gathering of dressed and costumed Sony AIBO robots appearing at Robosquare in Fukuoka City, Japan.
The act of playing with surfaces to explore identity and social/cultural relationships is something we begin to do from childhood, and it becomes an integral means of understanding our social and cultural values. The 19th century study of dolls in the Journal of Genetic Psychology consisted of anecdotal responses and ethnographic accounts of various cultures that practice some form of dollification of objects. The authors found that clothing and accessories were a large part of this play. In a period before the widespread mass production of toys and dolls, the reported accounts documented children dollifying a variety of objects that bared little if any resemblance to humans; among these were pillows tied with string around the middle, clothespins, flowers, bottles, sticks and chickens[Gra97]. Clothing, as an additional surface, can help transform the identities of these mundane objects and animate them into characters and playmates. The clothing of course can do the same for robots.
As part of our human experience, we play with surfaces both on ourselves and on objects as means to explore and understand the identity of social and cultural constructs. Robots of course are not immune from this human instinct and so we find quite a bit of anecdotal evidence of people, including roboticists, dressing their robots, Figure 1-2. The addition of robotic attributes to surfaces, through electronic-textiles and fabrics, presents an innovative opportunity to both integrate surfaces into robotics further and expand on the influence of clothing and surface on identity, not just to assist in socially and culturally situating robotics but for situating ourselves as well.
The Presentation of Self in Everyday Life is a seminal sociology book by Erving Goffman in which he formulated dramaturgical sociology. By dramaturgical, it is meant that our social interactions have a lot in common with theatrical performances in that we, as social-actors, attempt to influence how others perceive us presenting our identity by controlling our settings, manners and appearance[Gof59]. It suggests we might perform differently for different audiences, i.e. friends, family or work colleagues, and as such we’d desire to control our appearance by, for example, altering our style by controlling our surfaces, that is our clothes. It is from this perspective we consider fashion and clothing as costuming for identity. Soft-architecture robotic surfaces or as an extension, robotic clothing, offers us an augmented method for control of surface. Precisely how this control will socially perform or be perceived is yet to be fully explored.
Shifting Methodologies for Robotics
Among all the progress and developments in the field of robotics, there are two primary changes taking place that will help drive the fusion of fashion theory and robotics. The first change, as we saw in the anecdote about the QB Avatar telepresence robot, is that robots are entering the public space and joining our everyday social lives. The second change comes from technological developments, the core technology elements of robotics, i.e. sensing, actuation and computation, which are moving into soft and flexible electronic-textiles and fiber-electronic devices. Both of these changes will of course alter the aesthetic of robotics and overall require a shift or rethinking of the design methodologies currently employed in robotics.
The field of robotics has expanded beyond the domains of computer science and mechanical and electrical engineering. As robotics have transitioned from industrial manufacturing and laboratory research projects into the consumer market and thus into our homes and personal lives, the fields of design and psychology have gained more prominence in robotics, particularly if the robot is to be considered as a peer or companion. The robot design space is more complex for these social robots. Carl DiSalvo stated in observations on consumer robotics:
Because of the newness of robotics and the public's unfamiliarity with robots, the visual form of the robot often takes precedence in shaping our expectations of the robot and how we interact with the product.
-Carl DiSalvo [Car]
DiSalvo emphasizes the point that robots are not yet a daily experience for most people and so the lack of experience does not give the public a basis for forming any impressions. Instead the visual appearance of the robot and its aesthetics will be the driving force for initial public impressions which of course will shape our interactions. Initial impressions during social interactions make widespread use of nonverbal communications, and cues can signal much information. Clothing/fashion researchers have found that sartorial cues can play a fundamental role during initial impressions[Pau69][Kar11]. Clothing is a principle means by which we identify ourselves in the public space. It contributes to our identities by signaling, for example, our occupation, regional identities, gender, religion and social status to others via our chosen style of dress. As an illustration of this, clothing masculinity was found to significantly influence the perception of management characteristics of female applicants during a job interview. Those wearing masculine clothing were perceived as more forceful and aggressive[For88]. For robots, even the general physical appearance of an artificial agent can influence our perception of the information it supplies us as being fact or opinion [Forwn]. In this study, an artificial agent with a machine-like aesthetic resulted in a factual impression of the provided information. Conversely, the artificial agent with a more human-like aesthetic was considered to be offering an opinion. Even representations of gender and ethnicity in computer/machine agents have been shown to influence interactions[Lee08] [Pra07][Nas05]. It is foreseeable to utilize sartorial treatments or surfaces to assist in the design of social robots to help define their social roles among us and shape our interactions with them. Therefore by clothing we mean not just its morphology, we mean an entire system of clothing which incorporates the mechanics as well as the social semiotics of clothing and the performance aspects of style and fashion.
The question raised by current technological developments is what will the consequences of a textile and fiber morphology for electronics mean for the field of robotics? Electronic-textiles and fiber-electronics impart electronic sensing, actuation and computation into the fibers and textiles themselves. This will result in textile-based soft-architecture robotics which will change the face of robotic design. Traditionally, robots are generally constructed from a variety of rigid materials and surfaces often resulting in machine-like aesthetics and materiality. The emerging field of electronic-textiles and fiber-electronics represents a shift in morphology from hard and rigid mechatronic components towards a soft-architecture and more specifically, the possibility of a flexible planar surface morphology2. This morphology shift will open robotics to other disciplines which traditionally work with textiles and fabrics, like apparel designers and textile designers as well as textile manufacturing processes. The aesthetic of robotics will need to be reconsidered given these unique parameters. In order to try to understand this morphology transition we need to identify and build metaphors from more familiar classifications. Clothing is perhaps the most recognizable form of soft-architecture textile object and is nearly universal in its adoption and use. Therefore we must investigate sartorial materials, sartorial processes and semiotics of clothing design to inform electronic-textile applications in robotics. Understanding the mechanics of clothing and its functional aspect will be critical in working with robotic surfaces and soft-architecture robotics, but for social robotics there is a principal concern of social expectations associated with these designs, the semiotics of fashion.
The research presented in this thesis intends to address both these methodological shifts by developing a design framework that incorporates technological developments of electronic-textiles, apparel, textile design, robotics and fashion theory. This work intends to establish clothing and fashion as an untapped resource for robotics.
So how do we define Sartorial Robotics? Sartorial Robotics describes a practice of developing robotic surfaces which are situated between a functional technological skin that offers enhanced capabilities and the semiotics of clothing, i.e. fashion, which helps facilitate social play with identity. Both are combined merging methodologies of apparel and textile design with methodologies of robotics. This dissertation presents several demonstration projects built to support this hypothesis in fusing fashion theory and robotics. This research is meant to explore soft-architecture robotics in terms of human-robot social interaction. Precisely how this control will socially perform or be perceived is yet to be fully explored.