Reinventing the Cube: An Alternative Agent Design for

Shape-shifting Technology

Dr. Helen Hasenfuss

Interaction Design Centre, University of Limerick, Castletroy, Limerick, Ireland

Keywords: Tangible, Shape-shifting, Interactive Interfaces, Self-assembling, Multiagent Systems.

Abstract: In 2002 William Butera suggested the concept of a paintable computer. It is a blend of the concept of the

internet of things (IoT) and multiagent computing. He describes agents that function on a basic level, ideally

suspended in a liquid, so that they can be dispersed evenly but also be used to enhance the surface onto which

they are painted (Butera, 2002). Expanding on this concept to create a liquid computer that could change its

physical shape, i.e. it would be able to create unique 3D structures depending on user requirements. Such an

interface would be in a quasi-liquid state when inactive and could become solid when in use, comparable to

a Non-Newtonian fluid. This paper details an agent design that is orientated towards this kind of shape-shifting

interface technology.

1 INTRODUCTION

The concept of shape-shifting technology that can

morph into any user defined shape or form is a

growing field of interest. Inspiration is often taken

from nature and science fiction. Creating this kind of

technology requires ingenuity, creative problem

solving and a multidisciplinary approach. The aim of

this paper is to present and discuss a physical agent

design that is situated in the discipline of tangible,

shape-shifting interfaces.

The general consensus and portrayal of shape-

shifting technology is, that it is comprised of many

small parts operating together to create a larger entity.

This is essentially a multiagent system (MAS). These

small parts are represented by the term agents

throughout this paper. An agent is an autonomous

entity that can be represented organically as well as

inorganically and has a physical shape and form. It

has the capacity to make decisions and learn

according to its capability, based on behavioural

coding. There are numerous examples of biological

MAS that have received much attention in academia,

e.g. bee, ant or termite hives. The primary focus of

this research has been on A) communication

techniques (agent-to-agent or multiagent in terms of

sequencing, transmitting, receiving, deciphering,

error handling, etc) and B) the self-assembly process

(structural cohesion, attachment mechanisms, group

https://orcid.org/0000-1111-2222-3333

and environment cuing and orientation, error and

unforeseen event-handling, etc). Whilst the value of

this research is evident in the dissemination of this

knowledge into other disciplines, there has been little

progress with respect to the physical agent design

itself, i.e. the agent body. The cube is still the shape

of choice for many projects exploring self-assembly

(Gilpin et al., 2010, Romanishin et al., 2015, Roudaut

et al., 2016). Whilst the reasons for choosing the cube

are logical,

 6 facets,

 90degree dihedral angle,

 scalable,

 calculable

 familiarity in society (cuboid shape = building

blocks, Lego®, voxels, Minecraft),

it may not be the optimal shape on which to design

self-assembling or communication techniques. The

agent design discussed throughout this paper is based

on a PhD that was successfully completed in 2018

(Hasenfuss, 2018).

Section two will present related work in the field

of shape-shifting, tangible interfaces, also referring to

Tangible interface classification scheme.

The third section, whilst briefly indicating which

methodology is used, primarily describes the key

aspects of research on which the agent design is

Hasenfuss, H.

Reinventing the Cube: An Alternative Agent Design for Shape-shifting Technology.

DOI: 10.5220/0008116500150027

In Proceedings of the 3rd International Conference on Computer-Human Interaction Research and Applications (CHIRA 2019), pages 15-27

ISBN: 978-989-758-376-6

based. These aspects include the differentiation

between dynamic and static self-assembly, and

biological system of Solenopsis Invicta (fire ants).

The agent design, the Dod, is described in section

four. This encompasses its physical structure and

mechanisms. The characteristics that helped define

the Dod’s design are also presented.

The fifth section reflects on the how the Dod

fulfils the current design characteristics but is also

further adaptable in order to consider elements that

have yet to be resolved, such as energy manipulation:

its generation and utilisation within a multiagent

system. Following from this discussion, a brief

summation of the elements that define the Dod’s

uniqueness concludes the section.

2 RELATED WORK

According to the Tangible Interface (TI)

classification scheme as define by Ishii and Ullmer,

the agent design proposed in this study falls into the

category of constructed assemblies and to a degree,

of continuous plastic tangible user interfaces (TUIs)

(Ishii and Ullmer, 1997).

 Constructed assemblies are comprised of modular

components that can be arranged or organised into

larger more complex structures. A priority for

these assemblies is the manner of fit between

components or agents, their ability to create

assemblies in 3D or 3D relief and how these

assemblies interact with the environment (e.g

Topobo (Raffle et al., 2004)).

 Continuous plastic TUIs were developed in order

to provide the users with a malleable and flexible

building material with which to manipulate digital

information. It aims to accommodate user’s free

and direct interaction and not limit them to using

predefined, form-fixed elements in the interaction

process (e.g. clay (Ishii et al., 2004)).

Within the scope of the agent design described in the

original study (Hasenfuss, 2018), it is not only the

overall or whole entity that is changing (i.e. the larger

entity made up of smaller parts) but each individual

agent can change its form and size as well. This

behaviour creates a sub category to continuous plastic

TUIs. The primary function of the agent’s shape and

size change is to facilitate the generation of micro

structures to aid in overall assembly, e.g. line, curve

or cluster, to eventually create a complex interface,

like a mixing desk. The effect of this behaviour on the

macro structure can utilise the full potential of the

haptic modality. If each agent can change or react to

specific stimuli without effecting the overall

structure, it could create texturally dynamic and

fluidic interfaces. Similar behaviours are evident in

SMA or push-pin based interfaces (Minuto et al.,

2012, Follmer et al., 2013, Rozin, 2015). These types

of interfaces combine programmable matter concepts

with ambient computing, generating interfaces that

are more reactive to the individual user or their

environment. The ability of an agent to generate

structures as well as facilitate a mechanism for it to

physically react to its environment brings the agent

closer to emulating the unique haptic characteristic of

bi-directionality.

Haptics is essential in forming a holistic 3D

impression of the world (MacLean, 2008). Unique to

haptics is the opportunity to develop morphing or

shape-changing designs (Horev, 2006). Similar to bi-

directionality, malleability and reversibility are

characteristics that provide unique human-computer

interactions as well as design challenges (Coelho and

Maes, 2008).

Similar to the other senses, haptics also aims to

facilitate practical as well as emotional aspects of

interaction (Parkes and Ishii, 2010, Rasmussen et al.,

2012). A key difference in the design of future haptic

interfaces, is the concept of dynamic design. The

original design would no longer be static, like a

mouse or keyboard. Instead significantly more

ownership is handed to the user with respect to

creating malleable TUIs, suited to their specific

needs. Pushpin computing, smart fluid interfaces and

existing physical MAS projects are explored in the

following section.

2.1 Pushpin Computing

The concept of these interfaces is that each element

showing on the surface is individually actuated so that

it can move on designated planes. These elements can

consist of,

 actual pins, which are exposed (Poupyrev et al.,

2004, Follmer et al., 2013),

 pins that are covered with a flexible membrane

often fabric based (Iwata et al., 2001, Marquardt

et al. 2009, Leithinger and Ishii 2010),

 elements that are made of smart memory alloys

(SMA).

These interfaces focus on 3D relief, they are useful

for small-scale projects as there are no mechanical

mechanisms in the elements themselves and they are

capable of repeatedly transitioning between two

states (Coelho and Maes, 2008). These materials

allow for a greater degree of freedom with respect to

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

moving along the x, y and z-plane (Poupyrev et al.,

2004, Minuto et al., 2012) and attempt to create a

greater boundary transparency between the physical

world and digital information (Nijholt et al., 2012).

In pushpin computing, the moveable elements are

most often controlled individually. However, it is also

possible to program set behaviours in which several

elements may be grouped and move together. The

resolution, i.e. number of elements on display, can

vary to represent different structures as well as fulfil

specific requirements. Adjusting this resolution by

adding or removing pins / elements is often quite

difficult due to a complex pin-linkage-actuator-

computer setup. The overall setup confirms that

haptic devices are often difficult to implement, as

they are still bulky and rather unrefined when

compared to graphical user interfaces (GUIs).

Interfaces based on the pushpin technique are very

effective in achieving varied topological changes in

real-time and initiating new approaches to 3D

interactions. They primarily employ tactile sensing

and work with visuals either projected onto the

surface, light emitting or through inherent movement

of the elements (Iwata et al., 2001, Poupyrev et al.,

2004, Follmer et al., 2013).

2.2 Smart Fluids

Non-Newtonian (NN) and smart fluids were the

starting points of research in the original study,

specifically those exhibiting shear thickening

behaviour. These fluids are colloid suspensions and

the most commonly known NN fluid is cornstarch

and water (Oobleck). Depending on the suspension

ratio, when this fluid is agitated it becomes semi-solid

around the agitation source and for the duration of the

agitation (Bi et al., 2011). The particles have the

ability to form structure by assembling into chains or

hydroclusters. Once this external force ceases the

structures dis-assemble and return to a liquid state.

Magneto- and electro-rheological smart fluids

behave in a similar fashion to NN fluids, the main

difference being that a shear-thickening behaviour is

induced via a change in a magnetic field and electric

field respectively. Research into these disciplines

informed the characteristics for the core structure of

an agent body: particle material and construction.

These characteristics include size, shape, surface

topology, quantity and dispersion (Fall et al., 2012).

Interfaces based on these materials exploit the

aesthetic behaviour of fluids whereby practical

application or implementation has a lower priority in

the design process. Ferrofluid is most often used in

these types of projects because of its unique and

unusual reaction to a magnetic field. This is

dependent on strength and location of the magnetic

field. Ferrofluid is expensive to manufacture and is

not suitable for direct interaction therefore it is

necessary to enclosed it in a container (Masson and

Mackay, 2009, Saddik et al., 2011) or in a flexible

pouch (Hook et al., 2009, Jansen, 2010, Koh et al.,

2011). Even though the push-back force of Ferrofluid

is very low, it primarily has a visual appeal and has

the advantage of being able to create smooth,

noiseless, aesthetic transitions between varieties of

surface topologies. The typical setup requires the

Ferrofluid to rest a certain distance above a 2D array

of electromagnets. The electromagnets can be

controlled by a variety of factors: sound, light,

movement, etc. (Hook et al., 2009, Masson and

Mackay, 2009).

An interesting quality of these materials is the

ability to return to the original state and to repeat the

process of shaping and resetting. Similar to designing

vibration patterns to emulate textures, jamming

provides a variety of force feedback patterns to

emulate different materials beyond the surface texture

(Kim, 2010). Similar to pushpin interfaces, the

peripheral equipment required to make these

interfaces function as desired is substantial. For

example, for surface & depth detection equipment

such as IR cameras, projectors, computers and power

supplies are required to translate the interpretations of

malleable behaviour into digital information.

2.3 Multiagent Systems (MAS)

A MAS is a freestanding system and they are

pervasive through nearly all levels of existence, e.g.

individual cells working together, a flock of birds, the

stock market, solar system, etc.

With respect to artificial MAS, these interfaces

consist of modular agents, i.e. they are structurally

identical to each other. Thereby the haptic component

is inherent in their makeup in either of two ways: A)

in the physical handling of each individual

component (e.g. a tile) (Gorbet et al., 1998) or B) in

the physical handling of the overall system created by

many small parts.

The unique quality of modular interfaces is the

ability to add and subtract agents (Gorbet et al., 1998,

Parkes and Ishii, 2010). This presents a unique

challenge with respect to communication between

each component. Distributed sensor networks (Patten

et al., 2001, Rekimoto, 2002, Lifton et al., 2004) or

self–organisation models are applied to create an

adaptable and flexible system. From a haptic

perspective, these systems have a greater emphasis on

Reinventing the Cube: An Alternative Agent Design for Shape-shifting Technology

kinaesthetics since the interfaces are completely 3D

(McElligott et al., 2002, Kim, 2010). The peripheral

technology, as described for pushpin and smart fluid

interfaces, is primarily intrinsic to each agent.

The emphasis of a MAS is on agent autonomy.

The qualities from the research gathered in these

disciplines and projects helps inform a set of agent

characteristics that will be incorporated into the final

agent body design.

3 METHODOLOGY

The methodology used in this project is the STEAM

approach. Whilst a detailed description of it will not

be given in this paper, it is sufficient to highlight that

this methodology is a combination of inductive and

deductive reasoning and utilises the diversity of the

STEAM disciplines. Creative and scientific

techniques are merged to create informed and viable

sources of research that can be explored further

(Hasenfuss et al., 2018).

An important characteristic of MAS agents is the

ability to self-assemble particularly with respect to

3D tangible interfaces. The term self-assembly more

so than self-organisation is used because creating 3D

structures is not automatically implied in the latter. It

conveys the aspect of packing and orientation of

agents within a system (Rubenstein et al., 2014, Le

Goc et al., 2016).

An important distinction in self-assembly is that it

does not have to have intent inherent in its meaning.

For example, ball magnets and a hive of ants present

two different styles of self-assembly. When ball

magnets enter each other’s magnetic field, they begin

to self-assemble and thereby self-organize according

to the appropriate polar configuration (north-to-

south). In this case there is no intent other than the

naturally existing magnetic field or end-goals that can

change according to a dynamic environment. In

contrast to the ball magnets, when examining the way

ants self-organise, it becomes apparent that they work

together to achieve an end-goal (creating a bridge,

raft, etc). It can be argued that ants only follow a

specific program of behaviour similar to magnets, but

a primary difference is a degree of choice.

The latter type of self-assembly is described as

intent-ful or dynamic self-assembly (Whitesides and

Grzybowski, 2002). It is desirable for MAS based

shape-shifting interfaces, because it could allow a

user to stipulate the end goal, e.g. interface design,

rather than become involved in directing each

individual agent. Each agent would have enough

autonomy to deal with the task and some unexpected

events, e.g. agents being pushed, removed, breaking,

etc (Rubenstein et al., 2014).

The ability to dis-assemble represents bi-

directionality, which is a core component to haptics

and tangible interaction. It indicates that a system is

flexible, reusable and programmable in real time. The

final design (the Dod) attempts to fulfil this

mechanism but it also aims to address issues that can

be problematic by the inherent qualities of current

computing technology. For example, if specific

components of a laptop are broken beyond repair, the

entire laptop is usually scrapped by the average user

as repairs are often too expensive. This approach is

wasteful and detrimental to the environment,

something society can no longer afford. Therefore,

designing technology that can be reusable, even if

several components or mechanisms fail, aims to help

change the attitude to technology whilst

simultaneously creating a more robust and efficient

system. For example, in a MAS based interface, even

if complete mechanical failure occurred in a few

agents, it should still be usable because of the agents’

structural shape, (more on this in section 4.1).

Considering issues such as the longevity of

technology, the ability of systems to self-repair

efficiently or the connection between technology and

user are important design considerations even though

they may seem at this point removed from the design

process.

Examining biological systems can provide

insights into developing a model by which it is

possible to program the behaviour to achieve this

intent-ful or dynamic self-assembly.

3.1 Biomimicry

Biomimetics had a strong influence in the research for

developing an agent body that would be able to

function as part of a larger system of agents. The most

efficient and successful self-organising and

assembling systems already exist in nature e.g. bee-

hives, ant and termites, etc. Each of these complex

colonies can have an excess of several thousand

living beings, however each individual has its own

task and is appropriately equipped to carry out this

task (Dumpert, 1978, Gordon, 2010). The manner in

which these insects communicate and interact and the

scale at which they exist provides valuable design

insights. Projects such as Bergbreiter’s mini jumping

robots (Bergbreiter,

2014) and microTug

(Christensen et al., 2015) represent progress with

respect to scaling mechanical system. However, wear,

stress and strain still limit the lifespan of inorganic

systems.

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

Advancing from NN fluids, a biological

equivalent was sourced and studied: Solenopsis

Invicta (fire ants). This species of ant was researched

for their ability to achieve specific structures

indicating that they could repeatedly assemble and

dis-assemble. Due to their natural habitat being prone

to flooding, these ants have adapted by being able to

build temporary rafts until they find a new area of

land. The raft consists entirely of ants, which

demonstrates that they can maintain structural

cohesion, buoyancy and can survive temporary

submergence in water. They also demonstrate

behaviour similar to NN fluids: when agitated, e.g.

swirled in a beaker, they maintain a semi-spherical

shape and when left alone begin to disperse again

over time, etc) (Kasade, 2014).

There are three methods as to how the ants hold

together: mandible-tarsus, tarsus-tarsus, and chemical

excretions through adhesive pads located at the end

of the tarsus (Mlot et al., 2011). With respect to self-

assembly the tarsus-tarsus connection appeared to

have the greatest potential. An ant tarsus ends with a

claw. The shape of this claw is of interest because it

is a bistable interlocking mechanism. The claw

cannot be retracted. Therefore, the arc is such that it

must be able to interlock with other ants but release

with ease simply due to a shift in position.

The aspects described above provided material for

initial agent design and prototypes. The process is

similar to the User Centered Design approach except

that the ‘user’ is replaced by the interplay between

scientific and artistic methodologies, i.e. the dialogue

between inductive and deductive reasoning. For

example, haptic exploration was an integral part to the

agent design process. It was accomplished through

the modelling and construction of prototypes.

Accuracy, functionality and practicality were primary

concerns, i.e. researching about various problems

(scaling, disassembly, reversibility, etc) that

previously existed in artificial MAS and trying to

improve or solve them.

With the addition of artistic exploration, a shift in

focus allowed for aesthetics, imagination and

creativity to take higher priority thereby increasing

the scope of the design. Whilst some of the

conclusions could have been attained through a

purely scientific approach, applying an artistic

methodology suited the projects progression because

on the opposite end of TUI development is how these

interfaces should be used or interacted with. This is

where human creativity and adaptability do not

necessarily fit into moulds predetermined by

scientific or logical approaches. More information on

this topic can be found in the original study

(Hasenfuss, 2018).

4 FINAL DESIGN: THE DOD

In this section the physical aspects of the final design

are discussed. Whilst the behavioural aspects were

also considered in the original study, they will not be

discussed in this paper. The word final may indicate

that it is the completed or finished article. Instead it

should be interpreted as being the final stage at this

point in the study’s development.

4.1 Agent Qualities

The following characteristics define the Dod’s

design. They are not listed in any order of importance

because each feature is interdependent on the other

and they will be discussed further in section 5.

1 A semi-spherical shape with an irregular, cratered

surface.

2 Non-hierarchical chain of command: autonomy to

function as individuals

3 The ability to morph: surface topology and

fundamental form

4 One material make-up and scalability – structural

affordances and inherent material qualities

5 Bi-directionality – the ability to assemble and dis-

assemble

6 Behavioural simplicity

An important distinction when using biomimicry

to help create artificial systems is how biological

elements that are adapted to suit the new application.

The difference between replication and emulation is

subtle, except that emulation also entails moving

beyond the original design, i.e. incorporating core

elements but having alternative or new functionality.

This difference is important because rather than

focusing on replicating already naturally existing

systems (ant, bee or termite hives, etc) it is necessary

to consider what a man-made MAS must be able to

do, e.g. building a makeshift raft versus interfaces

with specific functions and forms.

In the original study, the design process began

with a phase of replication. For the Dod design, a

museum artefact, a Roman dodecahedron, was

chosen as an alternative starting point to the cube see

Figure 1.

Replication allows for exploration of the original

artefact, its parameters and the variables that effect it.

As the characteristics mentioned above emerged, the

design was continuously adapted. This iterative

process, as well as allowing for the influence of art

and creativity, helped to gradually transition the

design from the replicative to the emulative phase.

Reinventing the Cube: An Alternative Agent Design for Shape-shifting Technology

Figure 1: Early ideas and prototypes demonstrate a

replicative phase.

4.2 Physical Description

The final design, the Dod, is based on a geometric,

Platonic solid: the regular dodecahedron. Platonic

solids are polyhedrons that have inherent symmetry

and the most commonly used platonic solid in

existing projects exploring multi agent self-assembly

is the cube. The dodecahedron is semi-spherical in

nature and harmoniously balances the curved fluidity

of nature and humans desire for controllable, accurate

geometry. A dodecahedron can flow or move freely,

within or outside of a potential transport medium,

whilst still offering support in a scaffolding capacity

due to its flat facets and defined edges. The alignment

of dodecahedra is such that a curved line can be

represented and created more easily than is possible

with the cube. This characteristic provides an

alternative perspective in relation to building 3D

structures (Sieden 1989).

The dodecahedron is used as the core structure but

is designed with the ability to extend arms via a

spiralling twist i.e. an outward rotation. This ensures

that the outer plates do not interfere with each other.

Rotation may also provide the option to be used as a

mechanism to aid in disassembly of each arm or the

reverse to act as a locking mechanism for self-

assembly, e.g. polymagnets (Polymagnet Correlated

Magnetics 2016). The arms are an inverted

pentagonal frustum whereby the small, narrow end is

located at the inner core dodecahedron. The full

extension consists of two full rotations of the outer

plate. When all arms are retracted the outer plates

meet forming another dodecahedron, essentially

creating two dodecahedra: the inner or core

dodecahedron and an external dodecahedron that

encases the inner one. The inner core can be either

solid or hollowed out – the computational parts would

be located here, see Figure 2a.

Magnetism is used to represent the mechanism of

self-assembly. 3mm ball magnets are inserted into

each of the 12 facets. Therefore, the current size of

the Dod is reliant on the minimum printable material

around the magnet when the arm is retracted, see

Figure 2b.

It is currently proposed that each arm would be

able to extend individually. However, depending on

material and degree of structural cohesion required it

may be necessary to have pre-configured Dods, either

physically or through a Dod’s behavioural coding

(e.g. assignment of relative and static arm IDs).

Figure 2: a) Hollow Dod and b) working Dod prototype.

Configuration in this instance refers to the number

and manner of arms extended. The arms can be either

arbitrarily chosen or predetermined, e.g. like an

octopus whose eight legs may have fixed neural

connection but are not locked to always preforming

the same task, see Figure 3.

Figure 3: (a) finding equilibrium on irregular, soft surface

topology (b) a top-heavy cluster held at the base by

tweezers.

5 DISCUSSION

The Dod will be analysed in relation to the previously

mentioned characteristics, indicating its viability as a

blueprint for an artificial agent design. Following this

are several scenarios that consider design adaptations

to accommodate one of the most challenging

obstacles in shape-shifting technology: the generation

and manipulation of energy.

5.1 Characteristic 1, 3 and 5

Through the addition of an extending arm

mechanism, it is possible for the Dod to transition

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

from semi-spherical to irregular shapes. It changes

the surface topology of the agent as well as affecting

the over MAS structure. It makes it possible for the

agent to be adaptable to a variety of environments

(controlled, natural, liquid, solid, etc).

An origami pentagon spring was used in order to

convey the rotational extension of the arm. Either

ends of the spring connect to the facets of the inner

dodecahedron and the outer plate respectively. For

convenience, in the majority of prototypes generated

throughout the original study, the spring formed a

vertical column. However, the spring was also tested

in the form folded as an inverted pentagonal frustrum.

The advantage of using this type of spring is that A)

origami is very versatile (form, structure, strength,

size, etc.), B) there is a wide choice of existing and

accessible materials, C) this spring design is

pentagonal in its fundamental structure and D) it is

scalable.

Origami is a relevant technique that is gaining a

growing research interest with respect to its ability to

change shape, structure and strength (Cheng et al.,

2014, Lv et al., 2014, Silverberg et al., 2014, Reis et

al., 2015, Overvelde et al. 2016). There are several

characteristics that make this art form very appealing

in the robotic and interface domain.

 Changeable surface topology,

 Ability to morph shape & form,

 Auxetic quality,

 Strength transformation through structural

rearrangement.

It can inform individual aspects of projects as well as

the being the primary focus of a systems design (Lv

et al., 2014, Silverberg et al., 2014, Hawkes et al.,

2015, Miyashita et al., 2015, Reis et al., 2015,

Malkinski and Eskandari, 2016). The materials used

to produce origami inspired designs are as diverse as

the artform itself, ranging from smart memory alloys

& polymers (e.g. polypyrrole PPy and gold (Liu et al.,

2003), temperature & light actuated materials (Corina

2014), electromagnetic waves (Liu et al., 2003,

Miyashita et al., 2015), bond or foil paper to even

dissolving completely when a task is completed

(Christensen et al., 2015). Its application is also

evident in Nano-technology in the form of micro-

origami. An advantage of origami is its ability to exist

as micro and macro structures and that its strength,

which is material related, remains relative to the size

of the structure.

5.2 Characteristic 2 and 6

Rather than having a central agent that controls the

entire system, it is sufficient to give agents local

knowledge of their own position and that relative to

their immediate neighbours (Le Goc et al., 2016).

With respect to information-transmitting techniques,

wireless and touch based connections have been

suggested in the research relating to MAS or swarm

computing communication. Planting seed agents

throughout a MAS is also a considered alternative

communication method. Seed agents do not need to

look or act different, but they may potentially have a

greater capacity to store information about localised

clusters that form and can communicate this

information to other seed agents creating a more

holistic interpretation of the macrostructure.

In relation to the Dod design, the possibility for

the agents to communicate via touch exists primarily

because of the manner in which the agents assemble,

i.e. facet-to-facet. Being able to extend arms also

allows the Dod to spread its sensory processing from

its inner core through its arms. This mechanism could

have the advantage of transmitting a variety of data,

e.g. the arms could convey tactile and thermal or light

data whereas the inner core could be used to sense

orientation and acceleration.

5.3 Characteristic 4

The current aim in research is to achieve an agent that

is 2-6mm in overall size (Gilpin et al., 2010). At this

scale, material from which the Dod may eventually be

made not only has an impact on its physical and

behavioural constitution, but also on how an interface

will be texturally perceived in its entirety. Texture is

an integral component of haptics. It is influenced by

material and its behaviour in the specific environment

as well as the physical shape of a structure. It can be

used to guide and engage users in establishing

interaction experiences.

Since the Dod is based on the dodecahedron

which has clearly defined facets, it is not the facet

surface that is designed to change, rather through the

extending arms, it is possible to alter the surface

topology of the overall dodecahedron shape. This is

achieved by defining specific arm configurations or

behaviours (e.g. form clusters, dense and smooth or

semi-permeable mass, etc). It greatly effects the

textural quality of each individual Dod and as a result

the overall MAS.

Early prototypes demonstrated that the

dodecahedron can be scaled to approx. 2mm in

diameter and still maintain its faceted structure, a

similar test was carried out on the current Dod design,

see Figure 4a. It was possible to print the reduced

Dods, with sufficient accuracy, with an Ultimaker 3D

Reinventing the Cube: An Alternative Agent Design for Shape-shifting Technology

printer. Removing the magnet made it possible to

scale down the design even further. These studies

primarily tested whether the structural stability of the

extended arm could be maintained. Only two states

were printed for this test (all arms retracted, and all

arms extended), see Figure 4b, as they represent

either extremes with respect to available

configurations. A Form1 3D printer was used for

these prints because of the higher degree of accuracy

possible. See Table 1 for measurements.

Figure 4: a) scaled PLA Dods (b) scaled resin Dods.

To date the most viable means to create micro or

Nano-sized agents is to construct them out of one

piece of material (Kummer et al., 2010) and with as

few complex, moving parts as possible, e.g. joints and

cogs, etc, as possible as these may suffer from strain

and wear. Considering these aspects, the Dod has the

potential to embody this design concept. The rotation

and twist of the arms is designed to be a smooth

motion that incurs minimal levels of friction and

resistance.

Table 1: Scaled Dods - The size of each agent is noted in

mm, according to its diameter.

Retracted

Confi

uration

Extended

Confi

uration

Lar

e 8.5mm 14mm

Medium 6mm 10mm

Small 5mm 8mm

5.4 Context of the Dod

As mentioned in the introduction, the primary shape

used in the development of agent communication and

self-assembly is the cube. Other shapes such as the

rhombic dodecahedron have been used (Bojinov et al.

2002), however achieving a balance between

complexity and controllability has been elusive to

date. Rather than place and direct each individual

agent within the system, it is more advantageous to

allow scope for unpredictability. The Dod design

encompasses the capacity for both of these elements.

Dodecahedrons do not fit together as completely as

cubes however therein lies the advantage. In

conjunction with the extending arm mechanism, the

formations (craters) that are created ensure that other

Dods can fit or lock into them. This closely resembles

the shear thickening behaviour of NN fluids but also

the type of physical inter-connectivity that biological

systems, such as ants, can achieve.

The concept of the extending arm mechanism as

a specific design feature is validated to a degree

through the work of Overvelde. In the field of

metamaterials, he developed an algorithm that can

calculate new complex structures based on the

concept of combining geometric polyhedra with a

prismatic extension from all or a number of specific

facets. These structures also utilise origami as a form

of manufacture (Overvelde et al., 2016, Overvelde et

al., 2017). The following table highlights the

similarities and potential for the Dod in the field of

metamaterials and in turn shape shifting technology.

Table 2: Comparison between Dod and metamaterials.

Overvelde’s

metamaterials

Dod design

Polyhedra

cube, hexagon,

tetrahedron,

octahedron

dodecahedron

Extrusion

rismatic conical

Rigidity

fixture of

specific facets to

maintain the

reconfigure-

ability of the

macrostructure

fixture of

specific facets to

maintain

stability of the

agent

Structure

Macro-Units

‘grow’ from

replications of

the

microstructure

Through self-

assembly of

multiple agents,

macrostructures

are create

5.5 Energy

The ability of an artificial agent to manipulate and

convert energy is a challenging design problem. The

ideal aim is that any artificial and autonomous agent

should have a mechanism that enables it to convert

energy and that it can complete this process

independent of external assistance (e.g. a user having

to change a battery). This is not only important for the

self-sufficiency of the agent, but it also affects other

processes such as scaling and freedom of movement.

Energy gradients are present throughout any system

that moves from a state of imbalance to equilibrium.

Since the flow from higher states to lower states

occurs naturally or the system has a tendency towards

balance, this process often requires little or no extra

energy input. The energy required is to reverse this

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

process so that it can begin again. By using the

environment or another secondary function to provide

this energy would be ideal, e.g. motion / friction to

provide electrical energy, the users (touch: galvanic

response), pressure, photosynthesis, heat, hydration,

chemical exchange via an electrolyte, etc. Taking

these considerations into account also guided the

design process.

The following section describes several energy

processes. Some are linked to possible mechanisms

of motion, that were useful in developing scenarios as

to how the Dod design could be adapted in future

research.

5.5.1 Organic Process

Organic energy processes are of interest in relation to

agents that could eventually be 3D printed or grown

from biological material. The osmotic process, which

utilizes energy gradients is of interest in conjunction

with a Dod that is constructed of semipermeable

materials. The osmotic process is reliant on semi

permeable membranes in order to enable energy

gradients to be formed, e.g. sodium potassium pump.

The interesting feature is that this kind of gradient

enables the ions to move against the concentration

gradient, i.e. move from lower to higher

concentrations. For this scenario, the Dod’s

environment is liquid based and the process of

extending an arm is considered as the structure in

which to setup the appropriate gradient. The origami

spring is the equivalent to the vessel and is made of a

semi-permeable membrane, whilst the space between

the spring and the outer skin represents a filled body.

If the fluid in the spring has a high osmotic value, less

energy is required to create a higher concentration

gradient and the fluid in the body contains small

components that can cross past the semi-permeable

membrane. This osmotic value will attract volume out

of the body thereby by expanding and simultaneously

untwisting the spring as the osmotic gradient is

equalized. Expansion of the arm would occur

naturally due the higher osmotic value in the spring.

The arm retracts when energy is applied and the

smaller components, including fluids, are transported

back into the body reducing volume with in the spring

thereby causing it to twist and retract.

Algae and lichen are interesting organisms not

only because of their ability to photosynthesise but

also because of their symbiotic relationship. Lichen

are robust and adaptive to their environment. They are

made of two components: a fungus and algae and / or

cyanobacteria (photobionts) (British Lichen Society,

2018). The photobionts are capable of photosynthesis

whereas the fungus component acquires food through

its immediate environment. This latter component is

responsible for the vast diversity of lichen, i.e. the

variety of minerals and materials on which the fungus

feeds.

Yet another variant of the concept of symbiotic

relationships is one that was also an initial starting

point for this project: combining living matter with an

inorganic fluid that could exhibit shear-thickening

qualities. In the case of active fluids, swimming

bacteria are introduced into a lyotropic liquid crystal.

Among the valuable results to emerge from this

project, a key development was using the bacteria as

an energy source thereby creating and internal power

supply for the liquid crystal (Zhou et al., 2014)

The primary issue with natural or organic

processes is Time. These processes have taken time

to develop and perfect. In contrast, in the computing

industry, it is not often about achieving the most

efficient system, instead it is about achieving the

maximum that materials can offer. Rather than

propagate a throwaway culture through constant new

updates, or releases, MAS based interfaces may make

it possible to create a highly individualised computer

that acts as an extension or augmentation of human

abilities but that takes a few hours to assemble, as

opposed to a few seconds. For example, it takes MITs

new living tattoos 12hrs to react to chemical stimuli,

(Liu et al., 2017).

5.5.2 Triboelectric Energy

Wearable technology is a domain in which finding

power supplies other than traditional batteries is very

important - not only for aesthetics but also for

practicality. Whilst many devices are still self-

contained (bracelets, watches, etc) the development

of smart fabrics requires an alternative power supply,

e.g. solar energy - copper filament or optical fibres

can be woven into fabrics and materials such as

perovskite can be constructed in a chainmail fashion

which can be used as fabric (Hu et al., 2017). A

branch of wearable technology in which finding less

bulky power supplies is of great importance is active

prosthetics.

The development of smart skin aims to have more

realistic looking prosthetics but also to aid in the

generation of sufficient electricity to help power

certain aspects of a prosthetic. This is done through

research into the triboelectric mechanism (Rekimoto,

2002). It is based on the piezoelectric principle and

generates energy from friction forces as parts of the

prosthetic comes in contact with different materials.

The triboelectric effect is being coupled with forces

Reinventing the Cube: An Alternative Agent Design for Shape-shifting Technology

such low electric fields, sound, motion, light, etc. to

utilize environmental energy (Jeong et al., 2014, Cui

et al., 2015). In the attempt to replicate the human

skin structure and physiology, qualities such as

flexibility, elasticity, self-repair and identity are

useful concepts to potentially incorporate into MAS

agents. For example, just as a form of identity is

inbuilt into fingerprints it may be possible to

accomplish a similar feature for the Dod. In this way,

each Dod can be individually identified through an

exterior marker (unique skin pattern) and would not

have to send a digital tag as a means of identification.

The skin and what it represents, i.e. a self-renewing

boundary layer that has the potential to filter or absorb

specific materials, expand and contract, and to protect

from a wide variety of environments, is an interesting

approach particularly referring again to developing

the Dod from a biological perspective.

In this scenario the spring creates the core strand

of the extending arm, whilst the skin connects the

edges of the outer plate to the edges of the inner core.

If the material of the skin is designed like a

triboelectric generator then it has the potential to

generate sufficient energy to execute secondary

function such as extending and retracting the arms.

Another feature of skin is its galvanic response.

Several projects utilise this response as a means of

interaction with an interface (Rekimoto, 2002). This

has the advantage that an interface may be tuned to a

particular individual but also that the interface is only

re-activates when the user is touching it i.e. the user

provides the alternative energy source, similar to the

symbiotic nature of the lichen.

Currently the power generated by the galvanic

skin response is insufficient to power an entire

interface, and there are issues with the use of man-

made piezoelectric devices, e.g. packaging material

durability and device lifespan. Research into other

organic materials that generate piezoelectricity (bone,

tendons, DNA) indicates that it may be possible to tap

into other potentially renewable sources of energy

(Guerin, 2015). In conjunction with scaling the Dod

to the appropriate micro level may shift the size to

power ratio and make this source of piezoelectricity a

viable alternative.

5.5.3 Chemical Process and Motion

Alternatives to a biological reaction are systems that

process energy through chemical reactions. These

reactions can be quite potent, difficult to control and

are usually not bi-directional, i.e. once a reaction

occurs it is often difficult to retrieve or return to the

initial components (Bergbreiter, 2014, Miyashita et

al., 2015, Vilela et al., 2017). However, there are

advantages depending on the application and its

context. Similar to the living interfaces research

conducted in MIT, MAS agents could be

programmed to react when they encounter specific

chemicals (Liu et al., 2017). If the Dod is constructed

such that the arms are flexible then it may be possible

to design them so that when they come in contact with

a specific environment that they fulfil their function,

e.g. when no water is present the arms stay retracted

and when water is present the arms extend either

altogether or just the arms that sense a water source,

through an increase in humidity (e.g. hydrotropism).

Using the environment to trigger a chemical reaction

that produces energy, has been primarily

implemented in conjunction with agent motion.

Photo- and chemo-tactic motion is achieved by

manipulating photo or concentration gradients of

light sensitive chemicals or polymers. When these

gradients are localised around each agent,

diffusiophoresis enables the agents to exhibit motion

(Lozano et al., 2016, Vilela et al., 2017).

The ability to create smart polymers that can

exhibit the behaviours of a variety of materials, may

be a viable intermediary step with respect to creating

a completely biological Dod. Hydrogels exhibit a

reversible behaviour which is advantageous even

though it is dependent on external conditions. Many

biomedical devices are designed to react to a specific

trigger within their environment. This is most evident

in the development of targeted drug delivery systems

(Leong et al., 2009, Palleau et al., 2013, Breger et al.,

2015).

Lastly, developments in the discipline of soft

robotics and microfluidics is of value with respect to

pneumatic motion. The Octobot is a project which

utilises the principle of pneumatics based on chemical

reactions. The previous particle design explored the

direct reaction with the environment, whereas this

design is analogous to a diesel engine. The reaction is

inbuilt (chemical reaction), and the resulting output

(gas/pressure) is used throughout the system (Wehner

et al., 2016). The system is controlled via a

microfluidic logic circuit which enables this design to

be considered for soft robotic applications where

shape shifting, and flexibility are key requirements.

The challenge in this system is maintaining the correct

balance between fuel injection, actuation and venting

pressure and exhaust rate. Translating this concept to

the Dod could have substantial potential as the pressure

generated in such a system could be used to extend and

retract the arms. The current version of the Octobot is

approx. the size of an SD card (11x15mm) which is

another positive attribute with respect to scaling agent

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

designs. The microfluidic circuit has demonstrated that

it is possible to create soft circuitry using alternatives

to electricity (Jansen, 1990, Wehner et al., 2016).

These projects demonstrate some of the alternative

mechanisms by which agents can function and

appropriate the energy available to them in the

environment. Whilst a large proportion of these

processes are dependent on external actuation, it is still

possible to learn and further influence the Dod’s design

by considering these possibilities.

6 CONCLUSION

The Dod is only one part of a larger study into MAS

agent design. Rather than create another prototype in

the domain of shape-shifting technology that cannot be

applied beyond the lab environment, the aim of the

study was to produce a viable blueprint. This blueprint

can act as the basis for developments in the present but

also provide a canvas for future adaptations once

applications and materials are better defined. This

paper has focused on the practical description, of an

agent, more so than the STEAM methodology that was

used to achieve and support the final design. However,

one aspect of the methodology that is evident in this

paper, is the diversity of disciplines that informed the

design and the process itself.

ACKNOWLEDGEMENTS

I would like to thank my supervisor Dr. Mikael

Fernstrӧm for his guidance and support throughout my

study. Thanks also go to the Irish Research Council for

funding the first 3 years of this study (Project ID:

GOIPG/2013/351).

REFERENCES

Bergbreiter, S. (2014) Why I make robots the size of a grain

of rice [online], TED Talks, available:

https://www.ted.com/talks/sarah_bergbreiter_why_i_ma

ke_robots_the_size_of_a_grain_of_rice?language=en

[accessed 7 September 2015].

Bi, D., Zhang, J., Chakraborty, B. and Behringer, R. P. (2011)

'Jamming by shear', Nature, 480(7377), 355-358.

Bojinov, H., Casal, A. and Hogg, T. (2002) 'Multiagent

control of self-reconfigurable robots', Artificial

Intelligence, 142.

Breger, J. C., Yoon, C., Xiao, R., Kwag, H. R., Wang, M. O.,

Fisher, J. P., Nguyen, T. D. and Gracias, D. H. (2015)

'Self-Folding Thermo-Magnetically Responsive Soft

Microgrippers', ACS Applied Materials & Interfaces,

7(5), 3398-3405.

British Lichen Society (2018) What is a lichen? [online],

available:http://www.britishlichensociety.org.uk/about-

lichens/what-is-a-lichen [accessed Jan 2018].

Butera, W. J. (2002) Painting the Computer [Doctorate of

Philosophy in Media Arts and Sciences], unpublished

thesis (Doctorate of Philosophy), Massachusetts Institute

of Technology.

Cheng, N. G., Gopinath, A., Wang, L., Iagnemma, K. and

Hosoi, A. E. (2014) 'Thermally Tunable, Self-Healing

Composites for Soft Robotic Applications',

Macromolecular Materials and Engineering, 299(11),

1279-1284.

Christensen, D. L., Hawkes, E. W., Suresh, S. A.,

Ladenheim, K. and Cutkosky, M. R. (2015)

'μTugs: Enabling microrobots to deliver macro

forces with controllable adhesives', in 2015 IEEE

International Conference on Robotics and Automation

(ICRA), 26-30 May 2015, 4048-4055.

Coelho, M. and Maes, P. (2008) 'Sprout I/O: a texturally rich

interface', in Proceedings of the 2nd international

conference on Tangible and embedded interaction,

Bonn, Germany, ACM, 221 - 222.

Corina (2014) Heat, Light and Sound Reactive Materials

[online], available: http://scin.co.uk/heat-light-and-

sound-reactive-materials/ [accessed 10 Jan 2018].

Cui, N., Gu, L., Liu, J., Bai, S., Qiu, J., Fu, J., Kou, X., Liu,

H., Qin, Y. and Wang, Z. L. (2015) 'High performance

sound driven triboelectric nanogenerator for harvesting

noise energy', Nano Energy, 15, 321-328.

Dumpert, K. (1978) The social biology of ants, Bath: The

Pitman Press.

Fall, A., Bertrand, F., Ovarlez, G. and Bonn, D. (2012) 'Shear

thickening of cornstarch suspensions', Journal of

Rheology.

Follmer, S., Leithinger, D., Olwal, A., Hogge, A. and Ishii,

H. (2013) 'inFORM: Dynamic Physical Affordances and

Constraints through Shape and Object Actuation', ACM,

10.

Gilpin, K., Knaian, A. and Rus, D. (2010) 'Robot pebbles:

One centimeter modules for programmable matter

through self-disassembly', in 2010 IEEE International

Conference on Robotics and Automation, 3-7 May 2010,

2485-2492.

Gorbet, M. G., Orth, M. and Ishii, H. (1998) 'Triangles:

tangible interface for manipulation and exploration of

digital information topography', in Proceedings of the

SIGCHI Conference on Human Factors in Computing

Systems

, Los Angeles, California, USA, 274652: ACM

Press/Addison-Wesley Publishing Co., 49-56.

Gordon, D. M. (2010) 'Colonial Studies', Boston Review, 59

- 62.

Guerin, S. (2015) Decoding Electrocactive Organic

Materials using Solid State Physics, Internal University

of Limerick Report, unpublished.

Hasenfuss, H. (2018) A design exploration of an agent

template for multiagent systems (MAS) for shape shifting

tangible user interfaces., unpublished thesis (Doctorate

of Philosophy), Univesity of Limerick.

Reinventing the Cube: An Alternative Agent Design for Shape-shifting Technology

Hasenfuss, H., Fraifer, M., Kharel, S., Elmangoush, A.,

Ryan, A. and Elgenaidi, W. (2018) 'It Takes Two to

Tango : Merging Science and Creativity to Support

Continued Innovation in the IoT Domain', Advances in

Science, Technology and Engineering Systems Journal,

3(5), 82-91.

Hawkes, E. W., Christensen, D. L. and Cutkosky, M. R.

(2015) 'Vertical dry adhesive climbing with a

100× bodyweight payload', in 2015 IEEE

International Conference on Robotics and Automation

(ICRA), 26-30 May 2015, 3762-3769.

Hook, J., Taylor, S., Butler, A., Villar, N. and Izadi, S. (2009)

'A reconfigurable ferromagnetic input device', in

Proceedings of the 22nd annual ACM symposium on

User interface software and technology, Victoria, BC,

Canada, ACM, 51 - 54.

Horev, O. (2006) Talking to the hand. An exploration into

shape shifting objects and morphing interfaces [Thesis

Report], unpublished thesis Interaction Design Institute

Ivrea18 Oct 2016].

Hu, X., Huang, Z., Zhou, X., Li, P., Wang, Y., Huang, Z., Su,

M., Ren, W., Li, F., Li, M., Chen, Y. and Song, Y. (2017)

'Wearable Large-Scale Perovskite Solar-Power Source

via Nanocellular Scaffold', Advanced Materials, 29(42),

1703236-n/a.

Ishii, H., Ratti, C., Piper, B., Wang, Y., Biderman, A. and

Ben-Joesph, E. (2004) 'Bringing clay and sand into

digital design, Continuous Tangible User Interface', BT

Technology Journal, 22(4), 287 - 299.

Ishii, H. and Ullmer, B. (1997) 'Tangible Bits: Towards

Seamless Interfaces between People, Bits and Atoms', in

Proceedings of the ACM SIGCHI Conference on Human

factors in computing systems, Atlanta, Georgia, USA,

ACM, 234 - 241.

Iwata, H., Yano, H., Nakaizumi, F. and Kawamura, R. (2001)

'Project FEELEX: Adding Haptic Surface to Graphics',

in Proceedings of the 28th annual conference on

Computer graphics and interactive techniques, ACM,

469 - 476.

Jansen, T. (1990) Strandbeest [online], available:

http://strandbeest.com/ [accessed Jan 2018].

Jansen, Y. (2010) 'Mudpad: Fluid Haptics for Multitouch

Surfaces', in CHI 2010: Student Research Competition,

Atlanta, GA, USA, 4251 - 4356.

Jeong, C. K., Park, K.-I., Son, J. H., Hwang, G.-T., Lee, S.

H., Park, D. Y., Lee, H. E., Lee, H. K., Byun, M. and Lee,

K. J. (2014) 'Self-powered fully-flexible light-emitting

system enabled by flexible energy harvester', Energy &

Environmental Science, 7(12), 4035-4043.

Kasade, N. (2014) 'A ball of fire ants behaves like a material',

[Video online], available:

https://www.youtube.com/watch?v=0nzBaljEHj8

[accessed 17 Oct 2015].

Kim, H. (2010) 'Designing interactive kinetic surfaces for

everyday objects and environments', in Proceedings of

the fourth international conference on Tangible,

embedded, and embodied interaction, Cambridge,

Massachusetts, USA, ACM, 301 - 302.

Koh, J. T. K. V., Karunanayaka, K., Sepulveda, J., Tharakan,

M. J., Krishnan, M. and Cheok, A. D. (2011) 'Liquid

Interface: A malleable, transient, direct-touch interface',

ACM Computers in Entertainment, 2(7), 8.

Kummer, M. P., Abbott, J. J., Kratochvil, B. E., Borer, R.,

Sengul, A. and Nelson, B. J. (2010) 'OctoMag: An

Electromagnetic System for 5-DOF Wireless

Micromanipulation', IEEE Transactions on Robotics,

26(6), 12.

Le Goc, M., Kim, L., Parsaei, A., Fekete, J.-D., Dragicevic,

P. and Follmer, S. (2016) 'Zooids: Builidng blocks for

swarm user interfaces', in Proceedings of the Symposion

on User Interface Software and Technology (UIST), New

York, 97 - 109.

Leithinger, D. and Ishii, H. (2010) 'Relief: A Scalable

Actuated Shape Display', in Tangible, embedded and

embodied interaction, New York, NY, USA.

Leong, T. G., Randall, C. L., Benson, B. R., Bassik, N., Stern,

G. M., Gracias, D. H. and Heath, J. R. (2009) 'Tetherless

Thermobiochemically Actuated Microgrippers',

Proceedings of the National Academy of Sciences of the

United States of America, 106(3), 703-708.

Lifton, J. H., Broxton, M. and Paradiso, J. A. (2004)

'Distributed Sensor networks as sensate skin', BT

Technology Journal, 22(4), 32-44.

Liu, X., Yuk, H., Lin, S., Parada, G. A., Tang, T.-C., Tham,

E., de la Fuente-Nunez, C., Lu, T. K. and Zhao, X. (2017)

'3D Printing of Living Responsive Materials and

Devices', Advanced Materials, 1704821-n/a.

Liu, Y., Oh, L., Fanning, S., Shapiro, B. and Smela, E. (2003)

'Fabrication of folding microstructures actuated by

polypyrrole/gold bilayer', in TRANSDUCERS, Solid-

State Sensors, Actuators and Microsystems, 12th

International Conference on, 2003, 8-12 June 2003, 786-

789 vol.1.

Lozano, C., ten Hagen, B., Löwen, H. and Bechinger, C.

(2016) 'Phototaxis of synthetic microswimmers in optical

landscapes', 7, 12828.

Lv, C., Krishnaraju, D., Konjevod, G., Yu, H. and Jiang, H.

(2014) 'Origami based Mechanical Metamaterials',

Scientific Reports, 4, 5979.

MacLean, K. E. (2008) 'Haptic interaction design for

everyday interfaces' in Human factors and Ergonomics

Society.

Malkinski, L. and Eskandari, R. (2016) 'Magnetic Micro-

Origami' in Maaz, K., ed., Magnetic Materials, Rijeka:

InTech, Ch. 10.

Marquardt, N., Nacenta, M. A., Young, J. E., Carpendale, S.,

Greenberg, S. and Sharlin, E. (2009) 'The Haptic

Tabletop Puck: tactile feedback for interactive tabletops',

in Proceedings of the ACM International Conference on

Interactive Tabletops and Surfaces, Banff, Alberta,

Canada, ACM, 85 - 92.

Masson, N. and Mackay, W. E. (2009) 'WeMe: Seamless

Active and Passive Liquid Communication' in Jacko, J.,

ed., Human-Computer Interaction. Novel Interaction

Methods and TechniquesSpringer Berlin Heidelberg,

694-700.

McElligott, L., Dillon, M., Leydon, K., Richardson, B.,

Fernström, M. and Paradiso, J. A. (2002) '‘‘ForSe

FIElds’’ –– Force Sensors For Interactive Environments',

in Proceedings of the 4th international conference on

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

Ubiquitous Computing, Göteborg, Sweden, Springer-

Verlag, 168 - 175.

Minuto, A., Huisman, G. and Nijholt, A. (2012) 'Follow the

Grass: A Smart Material Interactive Pervasive Display' in

Herrlich, M., Malaka, R. and Masuch, M., eds.,

Entertainment Computing - ICEC 2012Springer Berlin

Heidelberg, 144-157.

Miyashita, S., Guitron, S., Ludersdorfer, M., Sung, C. R. and

Rus, D. (2015) 'An untethered miniature origami robot

that self-folds, walks, swims, and degrades', in 2015

IEEE International Conference on Robotics and

Automation (ICRA), 26-30 May 2015, 1490-1496.

Mlot, N. J., Tovey, C. A. and Hu, D. L. (2011) 'Fire ants self-

assemble into waterproof rafts to survive floods',

Proceedings of the National Academy of Sciences of the

United States of America, 108(19), 7669-7673.

Nijholt, A., Giusti, L., Marti, P. and Minuto, A. (2012)

'Proceedings of the 1st workshop on Smart Material

Interfaces: A Material Step to the Future', in Santa

Monica, California, ACM, 615 - 616.

Overvelde, J. T. B., de Jong, T. A., Shevchenko, Y., Becerra,

S. A., Whitesides, G. M., Weaver, J. C., Hoberman, C.

and Bertoldi, K. (2016) 'A three-dimensional actuated

origami-inspired transformable metamaterial with

multiple degrees of freedom', Nature communications, 7,

10929.

Overvelde, J. T. B., Weaver, J. C., Hoberman, C. and

Bertoldi, K. (2017) 'Rational design of reconfigurable

prismatic architected materials', Nature, 541(7637), 347-

352.

Palleau, E., Morales, D., Dickey, M. D. and Velev, O. D.

(2013) 'Reversible patterning and actuation of hydrogels

by electrically assisted ionoprinting', Nature

communications, 4, 2257.

Parkes, A. and Ishii, H. (2010) 'Bosu: A Physical

Programmable Design Tool for Transformability with

Soft Mechanics', in Proceedings of the 8th ACM

Conference on Designing Interactive Systems, Aarhus,

Denmark, ACM, 189 - 198.

Patten, J., Ishii, H., Hines, J. and Pangaro, G. (2001)

'Sensetable: a wireless object tracking platform for

tangible user interfaces', in Proceedings of the SIGCHI

Conference on Human Factors in Computing Systems,

Seattle, Washington, USA, ACM, 253 - 260.

Polymagnet Correlated Magnetics (2016) About

Polymagnets [online], available:

http://www.polymagnet.com/polymagnets/ [accessed 4

July 2016].

Poupyrev, I., Nashida, T., Maruyama, S., REkimoto, J. and

Yamaji, Y. (2004) 'Lumen: interactive visual and shape

display for calm computing', in SIGGRAPH '04

Emerging technologies, New York, USA.

Raffle, H. S., Parkes, A. J. and Ishii, H. (2004) 'Topobo: A

Constructive Assembly System with Kinetic Memory', in

Computer Human Interaction, Vienna, Austria, ACM.

Rasmussen, M. K., Pedersen, E. W., Petersen, M. G. and

Hornbaek, K. (2012) 'Shape-Changing Interfaces: A

Review of the Design Space and Open Reseach

Questions', in CHI, Hot Moves: Shape-Changing &

Thermal Interfaces, Austin, Texas, USA, 735 - 744.

Reis, P. M., López Jiménez, F. and Marthelot, J. (2015)

'Transforming architectures inspired by origami',

Proceedings of the National Academy of Sciences of the

United States of America, 112(40), 12234-12235.

Rekimoto, J. (2002) 'SmartSkin: an infrastructure for

freehand manipulation on interactive surfaces', in

Proceedings of the SIGCHI Conference on Human

Factors in Computing Systems, Minneapolis, Minnesota,

USA, ACM, 113 - 120.

Romanishin, J. W., Gilpin, K., Claici, S. and Rus, D. (2015)

'3D M-Blocks: Self-reconfiguring robots capable of

locomotion via pivoting in three dimensions', in 2015

IEEE International Conference on Robotics and

Automation (ICRA), 26-30 May 2015, 1925-1932.

Roudaut, A., Krusteva, D., McCoy, M., Karnik, A., Ramani,

K. and Subramanian, S. (2016) 'Cubimorph: Designing

modular interactive devices', in 2016 IEEE International

Conference on Robotics and Automation (ICRA), 16-21

May 2016, 3339-3345.

Rozin, D. (2015) Pom Pom Mirror, Bitforms Gallery, NYC.

Rubenstein, M., Cornejo, A. and Nagpal, R. (2014)

'Programmable self-assembly in a thousand-robot

swarm', Science, 345(6198), 795.

Saddik, A. E., Orozco, M., Eid, M. and Cha , J. (2011)

Haptics Technologies [online], available:

http://link.springer.com/book/10.1007%2F978-3-642-

22658-8 [accessed 20 september 2016].

Sieden, L. S. (1989) Buckminster Fuller's universe : his life

and work, Cambridge, Mass: Cambridge, Mass : Perseus.

Silverberg, J. L., Evans, A. A., McLeod, L., Hayward, R. C.,

Hull, T., Santangelo, C. D. and Cohen, I. (2014) 'Using

origami design principles to fold reprogrammable

mechanical metamaterials', Science, 345(6197), 647.

Vilela, D., Stanton, M. M., Parmar, J. and Sánchez, S. (2017)

'Microbots Decorated with Silver Nanoparticles Kill

Bacteria in Aqueous Media', ACS Applied Materials &

Interfaces, 9(27), 22093-22100.

Wehner, M., Truby, R. L., Fitzgerald, D. J., Mosadegh, B.,

Whitesides, G. M., Lewis, J. A. and Wood, R. J. (2016)

'An Integrated Design and Fabrication Strategy for

Entirely Soft, Autonomous Robots', Nature 536(7617),

451–455.

Whitesides, G. M. and Grzybowski, B. (2002) 'Self-

Assembly at All Scales', Science, 295(5564), 2418-2421.

Zhou, S., Sokolov, A., Lavrentovich, O. D. and Aranson, I.

S. (2014) 'Living liquid crystals', Proceedings of the

National Academy of Sciences, 111(4), 1265-1270.

Reinventing the Cube: An Alternative Agent Design for Shape-shifting Technology