Content uploaded by Eliot Graeff
Author content
All content in this area was uploaded by Eliot Graeff on Jul 24, 2019
Content may be subject to copyright.
Biomimetics, where are the biologists?
Final version of the Author’s Manuscript (AM)
Eliot Graeff, Nicolas Maranzana and Améziane Aoussat
DOI: https://doi.org/10.1080/09544828.2019.1642462
Engineering design, as the science framing the practice of design through the
elaboration of tools and processes, is constantly evolving towards new innovative
strategies. To thrive in their extremely competitive environment, it appears that
both industrial and natural worlds are highly dependent on innovation,
optimisation and selection. These commonalities have led designers to look to
living beings for inspiration. This innovation strategy, referred to as biomimetics,
isn’t a new approach but its methodological aspects are still under development.
This article deals with biologists’ contribution throughout the biomimetic design
process. After introducing the context and the experimental protocol, we
investigated the impact of possessing a background in biology during the practice
of biomimetics and compared our findings with experts’ opinion. The main idea of
this article is to show that to forego the integration of biologists is highly restrictive
and may be one of the reasons explaining the difficulties of implementing
biomimetics in the industrial context. Hence, this article argues for a new
methodological framework taking into account biologists, allowing biomimetic
teams to become truly interdisciplinary.
Keywords: biomimetics, descriptive models of the design process,
transdisciplinarity, design teams, biologists
Subject classification codes: include these here if the journal requires them
1. Introduction
Biomimetics is defined as “the interdisciplinary cooperation of biology and technology
or other fields of innovation with the goal of solving practical problems through the
function analysis of biological systems, their abstraction into models and the transfer into
and application of these models to the solution” (ISO/TC266 2015). Over the past
decades, biomimetics has established itself as one of the most promising strategies to
support innovative and environment-friendly products (Vincent et al. 2006; Hwang et al.
2015; Bar-Cohen 2006). Many studies have evaluated such processes and shown that they
allow a greater novelty of solutions and thus generate a larger amount of innovative
products (Keshwani et al. 2017; Ahmed-Kristensen, Christensen, and Lenau 2014).
However, biomimetic design processes keep facing a fundamental trade-off
between integrating biologists and finding ways to practice biomimetics without them.
Biologists potentially constitute both the greatest asset and the greatest challenge of
biomimetics. It has always been seen in interdisciplinary fields that the benefits of
diversity comes with strong barriers in terms of communication and understanding
between actors (Schöfer 2015; Schöfer et al. 2018). In the case of biomimetic design, a
quick review of the existing tools and processes raises the following question: where are
biologists? Most of the tools are designed to be used by engineers, all processes are meant
to be followed by engineers and the resulting feeling is that biologists aren’t considered
in the methodological equation of biomimetics.
On account of the field’s intrinsic need for biological knowledge, the current
strategy is to develop tools, such as databases of biological strategies (Deldin and
Schuknecht 2013; Nagel et al. 2010; Vincent et al. 2006). So far, this process works
thanks to the relative novelty of the approach and low number of practitioners. The
biggest database of biological strategies is AskNature which was created in 2008 and
gathers 1 671 biological strategies and 201 inspired ideas (Biomimicry Institute 2002).
Even if the database is enlarged little by little, there is a real need to hasten this process
in order for it to represent a viable solution in the long run (Vandevenne et al. 2014).
Biomimetic design teams that directly include a biological expertise (i.e. an actor from
natural sciences) may be the only coherent way to consider biomimetic as a long-term
design strategy (Lenau 2009; Snell-Rood 2016; Schöfer et al. 2018; Chirazi et al. 2019).
After the presentation of the context and the hypothesis, this article deals with the
following research question: Facing the current biomimetics’ methodological framework,
does knowledge in biology represent an asset for the practice of biomimetics? Through a
case study carried out on the insulation of container homes in Ivory Coast and the above-
presented biomimetic tool, AskNature, a comparative analysis of the results obtained by
students, with a background in engineering or in biology will be detailed. Then, they will
be compared with experts’ opinion collected through a survey. With the aim of bringing
out the fundamental question of the biologist's role in the biomimetic process, the
contribution of this article is to underline the need to considerate biologists during both
the practice of biomimetics and engineering design surrounding biomimetics.
2. Context
Through an overview of some common points between current design challenges and
life’s innovation mechanisms, this section addresses the question of the origin of the
biomimetic process and underlines its crucial need for a methodological framework.
2.1 Modern design challenges
Many consider the alignment of product design and sustainable development as one of
the greatest modern challenges (Bourgeois 2007; Sachs 2015). Indeed, whereas
innovation remains one of the main levers for competitiveness, decades of pressing
scientific warnings made governments and companies progressively realise that climate
change and its impact on our society will dictate the centuries to come (IPCC 2014). In a
world where the use of energy and raw materials will have to be minimised and where
waste stream will need to be drastically reduced, our ability to design entirely new
innovation strategies will be one of the keys to success.
Engineers often refer to design theories to solve design problems. Among these
theories, which outline the current framework of design practice, a significant number is
composed of problem-solving strategies that include abstraction processes (Lee, Pries-
Heje, and Baskerville 2011). This part of the process aims at considering the design
problem with a different perspective through its formalisation into theoretical models:
Universal Design Theory (Grabowski, Lossack, and El-Mejbri 1999), Synectics (Gordon
1961), TRIZ (Altshuller 1984), etc. Once modelized, the given set of constraints and
objectives are compared with those of solved cases from the same or other scopes of
practice. This analogical thinking can be extremely powerful but requires an important
source of resolved dilemmas to reach its full efficiency. The context above-mentioned,
leads us to look for a prism through which design teams will be able to generate
breakthrough innovations combining strong technical advantages and sustainability.
Since the survival of living beings depends on innovation guided by biological rules, the
biomimetic approach, as the extraction of biologic strategies to create analogic
innovation, can offer such a prism (Helfman Cohen and Reich 2016). Looking at the
strategies and adaptations of living beings appears to be a formidable learning tool for us
to consider, as an endless reservoir of solved puzzles (Bila-Deroussy 2015).
2.2 Bio-inspiration, a wish to learn from living beings
Bio-inspiration is not a new strategy. The living world was often an object of comparison
and a source of inspiration during the development of antic societies (Vincent et al. 2006;
Lenau 2009) and has led to several scientific and design breakthroughs throughout history
(Richter, Wells, and Kemp 2008; Eggermont 2015; Guerrero, Maestro, and Bottaro 2012;
Magyar et al. 2017), etc. From aeronautics (Bar-Cohen 2012) to the biomedical field
(Zhang 2012; Boeuf 2007), biomimetic success stories along with the dissemination work
of iconic figures, such as Benyus (Benyus 1997), granted bio-inspiration an international
recognition. This increasing interest of the global engineering community also uncovered
the lack of tools, processes and methods on bio-inspiration, leading engineers to formalise
a bio-inspired design process in order to systemise its use.
2.3 The need for the formalisation of biomimetic processes
The inherent difficulty of such an interdisciplinary and unframed field is firstly the
communication between the different actors. In 2015, to prevent any misunderstandings
researchers developed a strict linguistic definition of the most common, and often mixed
up, terms referring to the field: bio-inspiration, bionic, biomimicry and biomimetics
(ISO/TC266 2015). Based on these standards, this article will focus on biomimetics.
Biomimetic processes are based on two types of approaches. Either the design is
conceived after a biological discovery, leading to a new product often with a high added
value: the biology push approach (ISO/TC266 2015), or biomimetics is used as a
problem-solving process: the technology pull approach (ISO/TC266 2015). As the core
of engineering design in the industry, the formalisation of biomimetic methods and tools,
has been focusing on this problem-solving approach. A great number of methods have
been designed to help implement a technology pull approach, The procedural model of
doing bionics (Lindemann and Gramann 2004), Biomimetic design methodology (Lenau
2009), Problem-driven analogical process (Goel et al. 2014), etc. These methods can be
described with the same eight main steps (Fayemi et al. 2014). As a result, Fayemi
proposed a unified technology pull biomimetic process (Figure 1) which constitutes the
methodological framework of our study. Despite these methodological improvements,
the need for new tools soon emerged. Indeed, the lack of biological knowledge led
engineers to face a limitation with such processes: find and extract biological strategies,
functions, shapes, etc. A great number of tools have been conceived to deal with this
issue, leading to the creation of tool-based methods – Idea-Inspire process (Chakrabarti
et al. 2006), Biologically inspired concept generation (Nagel et al. 2010) – or to the
improvement of existing processes.
Figure 1. The unified technology-pull biomimetic process from Fayemi et al. 2017
Tools used in biomimetic design have mainly three origins (Fayemi 2016) :
• engineering, like 5-Whys (Ohno 1978) or Technical contradictions (Altshuller
1984),
• biology, like 16 Patterns of Nature (Hoagland and Dodson 1995) or Functional
modelization (Tinsley et al. 2007),
• biomimetics, like SAPPhIRE (Chakrabarti et al. 2006) or BIOTRIZ (Vincent et
al. 2006).
It has to be specified that the tools that are said to come from biology, are neither
biological tools nor tools designed to be used by biologists. They have been mainly
designed for engineers to learn about biological findings. Wanieck et al. (2017) performed
a classification and analysis of more than 40 tools. As often in engineering design, the
increasing number of tools leads the user into an unclear path as it becomes more
complicated to choose which tool to use, and under which conditions (Lahonde 2010).
This formalisation of the biomimetic process progressively highlights a
contradiction with both the theoretical and practical aspects: the need for biologists is
underlined in the literature but the methodological framework hasn’t been designed to
include these unusual profiles.
3. The biologist’s role in biomimetic processes
3.1 A paradoxical consensus leading to our research question
According to Wanieck et al. (Wanieck et al. 2017), the step of the biomimetic process
that holds the greatest numbers of tools is the step of identification of the biological model
(21 tools) leading the author to conclude ‘This fact also strengthens the necessary role of
biologists in the field of biomimetics’. Paradoxically, most of those tools are designed
by engineers, using engineering references and techniques and consequently appear to be
made for engineers, not for biologists. Thus, it becomes clear that the overall tendency is
to delay biologists' intervention or even to replace biologists by tools, without dealing
with their tricky integration into the biomimetic process. Therefore, tools are getting more
numerous and complex to compensate for the lack of biological knowledge, which leads
us to the research question of this article: Facing the current biomimetics’ methodological
framework, does knowledge in biology represent an asset for the practice of biomimetics?
As we’ve seen in the state of the art, some famous experts have been advocating for
decades for biologists’ inclusion. However, this inclusion isn’t the main focus of the
engineering design research on biomimetics. As a result, they are still poorly integrated
and largely considered as external resources while biomimetic processes and tools are
being designed. Thus, the aim of this article is to gather experimental data and experts’
point of view on biologists’ contribution during the practice of biomimetics. This
substantiated answer should help foster the research effort towards the integration of
profiles from natural sciences as a relevant improvement for biomimetic design teams
and processes. In order to discuss this topic, we focused on one of the most
comprehensive and complete biomimetic databases (Faludi 2017; Deldin and Schuknecht
2013) : AskNature (Biomimicry Institute 2002).
3.2 Hypothesis
This article deals with the contribution of biologists in the biomimetic process. Our
hypothesis is that having an expertise in biology increases the biomimetic process
efficiency. Indeed, we think that a background in biology will lead to a quicker, deeper
and more reliable search of biological strategies. Moreover, we expect tools like
AskNature to be used for different goals, depending on the user's level of biological
knowledge. Where, at a first level engineers can get inspired from living beings to
increase the innovation potential of creativity steps, a background in biology may lead to
a deeper, more comprehensive and more accurate understanding of biological strategies,
required for the abstraction and the direct transposition of biological mechanisms into
technology.
In order to test our assumptions, we designed a case study on the insulation of container
housing in Ivory Coast, and a survey to collect experts’ opinion.
4. Materials and methods
4.1 Materials
The case study presented in this article aims at developing an insulation for former
shipping containers so that they can be turned into comfortable houses in Ivory Coast.
Students had access to a short description of the case study and to basic data about Ivory
Coast and the context of the project: climatic information (type of climate, monthly
temperatures and precipitations), characteristics of the containers and of the planned
village of containers, expectation of the industrial actor in terms of aesthetic.
The first tool, the Biomimicry Taxonomy (Figure 2), was also distributed to the
participants. This open-access tool was created by the Biomimicry 3.8 Institute to
transpose technical problems from non-biological fields into requests suitable for
AskNature (Baumeister 2014). The Biomimicry Taxonomy is constituted of three levels:
• Group, which refers to the overall function (ex: protect from physical harm),
• Sub-Group, which refers to a specificity linked with the overall function (ex:
protect from living threats),
• Function, which refers to a more specific case in the sub-group (ex: protect from
Fungi).
This multi-scale taxonomy acts as a skeleton for AskNature and enables the user to
formulate requests using a common linguistic frame, directly connecting technical issues
with AskNature data. AskNature is an open-access database created by the Biomimicry
3.8 Institute (Biomimicry Institute 2002). It lists biological organisms, strategies,
functions and more generally characteristics. This data is then made readily available
through an intuitive search engine.
Figure 2. The Biomimicry Taxonomy (Baumeister 2012)
For purposes of accessibility to consequent samples, our experiment compares the results
obtained by individual students with a background in biology versus those obtained by
engineering students instead of actual professionals. The experiment has been conducted
with a total sample of forty-two students divided as explained in Table 1.
Table 1. Sample characteristics and distribution for the case study
Profiles (and scientific training)
Level of Education
Biomimetic
knowledge
Number of
students
Engineers (Industrial engineering, product
design, interaction design, creativity, etc.)
Master’s degree
(M2)
None
21
Biologists (Agronomy, molecular biology,
physiology, soils sciences, environment, etc.)
Master’s degree
(M1 & M2)
None
21
• Energy
• Gases
• Liquids
• Solids
• Gases • Energy
• Chemical entities
• Solids
• Liquids
• Chemical entities
• Gasses • Energy
• Org
a
nisms •
S
oli
ds
• Liquids
•
L
iving
ma
ter
ial
s
•
Non
-
l
iv
ing
mate
rials
• Catalyze breaking of bonds
• Organic compounds
•
Ino
r
g
an
i
c
c
o
m
poun
d
s
•
P
o
ly
m
e
r
s
•
Cle
a
ve
h
al
o
gens
f
r
om
o
rg
ani
c c
o
mpo
u
n
ds
• Cleave heavy metals f
r
om organic compounds
• Balance/orientation • Shape and pattern • Time and day length
• Sound and other vibrations • Temperature • Motion • Pain • Body awareness
• Touch
and mechanical forces •
Ch
e
micals (odor
,
taste, etc.) • Atmosp
her
ic
co
n
dit
i
o
n
s
• Light (visible spectrum) • Light (non-visible spectrum) • Electricity/magnetism
•
Disease
•
Differentiate si
g
nal from noise
•
Transduce/conve
r
t signals
• Resp
o
nd to
s
ignals
• Tacticle • Chemical (odor, taste, etc) • Vibratory • El
e
ctrical/magnetic
• Li
g
ht (vi
s
ible
s
pectr
u
m) • Light (non
-
visi
b
le sp
e
c
t
rum) • Sound
• Through air • Through liquid • Over land • Through
s
olids
• Catalyze making of b
ond
s •
On
de
m
an
d
• A
t
ta
c
h a fu
n
ctional
g
r
oup
•
D
e
tach a
f
u
n
c
ti
o
nal
g
rou
p
• Ino
r
g
a
n
ic
c
o
m
pounds
•
O
r
gani
c
c
ompo
u
nd
s
•
Spec
i
c s
t
er
e
oi
s
ome
r
s
• M
i
n
e
r
al
cry
sta
ls
• Polymers • Metal-based compounds • Molecula
r
devices
• Thermal energy • Radiant energy (light)
• Chemical energy • Mechanical energy
• Electrical energy • Magnetic energy
• Structure
• Self-replicate
•
Adapt beha
v
iors • Optimize spa
c
e/mate
r
ials
• Adapt genotype • Adapt phenotype • Coevolve
• Chemically generate ow of electrons
• Surface tension • pH • Solubility • Electron transport
• Oxidation state • E
l
ectric charge • Conductivity
• Che
m
i
c
al poten
t
ial
• React
i
vit
y
w
i
t
h
water
• Energy state • Free radical reactivity • Concentration
• Speed • Position
•
B
u
o
ya
n
cy •
L
ig
h
t/co
lo
r
•
Ma
terial
c
hara
cteris
t
i
c
s
• Size/shape/mass/volume • Pressure • Density • Phase
• Maintain biodiversity •
B
iological control of populations, pests, diseases
•
R
eg
ulate atmos
p
h
e
ri
c
composition
•
Re
g
ula
t
e
cl
im
at
e
• Disperse seeds
•
C
o
ntrol er
os
ion
a
nd sedimen
t
•
Reg
ul
at
e
w
a
ter stor
a
ge
• Cyc
l
e n
ut
r
i
ents
• G
e
ne
rate soil
/
re
n
ew
f
e
rti
lity
• D
et
o
xi
c
at
i
on/
puri
cat
ion
o
f
a
i
r/
w
ater
/
wa
st
e
•
M
anage disturbance in a community • Regu
l
a
t
e h
y
dro
l
ogical ows
• Pollinate
• Within a (eco)system • Cooperate/comp
e
te between (eco
)
system
s
• Within the same species • Cooperate/compete between different species
• Coordinate by self-organization • Activities • Systems
• Buckling • Deformation • Fatigue • Melting • Fracture/rupture
• R
epro
d
uc
t
i
on o
r g
ro
w
th
• Cellular processes
• Maintain homeostasis
• Cr
e
e
p
• Co
m
pr
es
sion
• Mechanical wear • Chemical wear
• Impact • Tension • Turbulence
• Shear
•
Thermal
s
h
o
ck
• Temperature • Nuclear
r
adiation
• Light • Chemicals • Fire • Ice
• Loss of gases • Dirt/solids
•
L
os
s
of
l
i
q
ui
d
s
•
Gase
s
• Ex
c
ess l
i
q
ui
ds
•
Wind
• Fungi • Microbes
• Animals • Plants
• In/through gases
• In/
o
n liquids
• In/on solids
• Temporarily
• Permanently
• Gases
• Liquids
• Solids
DISTRIBUTE
STORE
C
A
P
TU
R
E
,
A
B
S
O
RB,
OR
F
I
LT
E
R
PHY
S
I
CA
LLY BR
E
AK DO
W
N
CHEMICALLY BREAK DOWN
COMPUTE LEARN [EN/DE]CODE
SEN
SE
SI
G
NALS
/
ENVIRONMENTAL CUES
P
R
O
C
E
SS
SIGNAL
S
SEND SIGNALS
NAVIGATE
CHEMICALLY ASSEMBL
E
TRANSFORM/CONVERT ENERGY
PHYSICALLY ASSEMBLE
REPRODUCE
ADAPT / OPTIMIZE
MODIFY CHEMICAL/ELECTRICAL
STATE
MODIFY PHYSICAL STATE
PROVIDE ECOSYSTEM SERVICES
COOPERATE
COORDINATE
PREVENT STRUCTURAL FAILURE
REGULATE PHYSIOL. PROCESSES
MANAGE STRUCTURAL FORCES
PROTECT FROM NON-LIVING
THREATS
PROTECT FROM LIVING THREATS
MO
V
E
AT
TACH
EXPEL
MO
V
E O
R
STAY PUT
PRO
CES
S
INFORMATION
GET, STORE, OR
RESOURCES
DISTRIBUT
E
MAKE
MODIFY
MAINTAIN
COMMUNITY
PHYSICAL HARM
PROT
E
CT
F
RO
M
BREAK DOWN
© 2008-2017 Biomimicry Institute, v.6.1
Creative Commons BY-NC 4.0
FUNCTION
SUB-GROUP
GROUP
v. 6.1
BIOMIMICRY TAXONOMY
© 2008-2017 Biomimicry Institute, Creative Commons BY-NC 4.0 | Biomimicry.org
asknature.org
At each step of the process, participants had to report their findings on an answer sheet
composed of four parts:
• User’s profile
• Results and follow-up of the biomimetic procedure
• Tools' evaluation (Biomimicry Taxonomy and AskNature) on a 5-point Likert
scale. For each tool, five criteria, based on Nielsen’s criteria about usability
assessment (Nielsen 1993), were evaluated: satisfaction, learnability, error,
wealth of information (derived from Nielsen criterion ‘Efficiency’) and precision
of information (derived from Nielsen criterion ‘Efficiency’).
• Feedbacks from the users: What were the hardest steps you faced?
The design process used in the experiment is largely based on the unified technology pull
biomimetic process (Figure 1). Due to time constraints, two modifications were made:
steps 4 and 5 were gathered and students didn’t deal with the 8th step, leading to a 6-step
process (Figure 3).
Figure 3. Comparison of the steps composing the unified problem-driven biomimetic
process (Fayemi et al. 2017) and the process used in the experiment.
For the second part of the study, a survey was designed and submitted to 17 experts on
biomimetics. Nine of the experts came from France and the rest from a range of other
countries. Their background was biology (4), architecture (4), engineering (6) and other
(3). They were identified either because they took part in the scientific experts panel
present at the 2018 Biomim’Expo conference in Paris, or because they are members of
the CEEBIOS (Centre Européen d'Excellence en Biomimétisme de Senlis, the French
association which co-organizes the Biomim’Expo conference and supports the growth of
biomimicry in France) (Table 2).
Table 2. Sample characteristics and distribution for the survey
Profiles (N=17)
Researcher
Consultant
Both researcher and
Engineer
only (n=5)
only (n=4)
consultant (n=6)
only (n=2)
Master’s degree
0
2
2
1
PhD
5
2
3
1
Accreditation to supervise
research
0
0
1
0
The survey is twenty-five questions long but in the context of this article only two axes
will be presented: Are biologists required for the practice of biomimetic? And why
integrate a biologist in biomimetic design team?
4.2 Methods
The protocol of the experiment remains the same for each session.
• Presentation of biomimetics and of the case study (15 min)
• Presentation of the protocol, the tools (15 min)
• Problem analysis (step 1), abstraction of technical problems (step 2) and
transposition to biology (step 3) (30 min)
• Identification and selection of biological organisms or strategies on AskNature
(step 4) (30 min)
• Abstraction (step 5) and transposition (step 6) of the biological solution into a
technical one (20 min)
The authors then analysed users’ answer sheets. Firstly, we focused on the number of
determined functions (step 1), technical problems (step 2), requests (step 3) and
organisms (step 4) through a quantitative (t-test) and correlation analyses (Spearman test).
Because of the short duration of the experiment, the idea was not to actually find ideas
for containers insulation but rather to collect data in order to evaluate the level of
understanding of the users. Then, through the analysis of the abstracted strategies (step
5) and technical solutions (step 6), we tried to evaluate the understanding of biological
strategies. At first, we wanted to grade the users’ level of understanding of the biological
phenomena, but there appeared to be too many biases to get to a relevant result. As an
example, if a biologist talks about photosynthesis, even without any details, he knows
what the underlying phenomena are, where an engineer doesn’t. As a result, the level of
understanding appeared too hard to assess without additional data.
As a consequence, we decided to focus on the biological strategies that haven’t
been understood, which is much easier to assess. The respective amount of misunderstood
strategies will be presented for each population and compared with a Chi2 test.
Furthermore, in order to help the reader, the compared results are always
presented in the text formalised as follows: (biologists results vs engineers results, type
of test, p-value:).
This analysis led us to consider a set of main tendencies that will be detailed in
the next section and were used to formulate the possible answers during the design of the
survey.
5. Results
We will look at the results obtained at the scale of the general process before analysing
the results in more depth through a step by step analysis.
5.1 Results on the global process
For each step of the process, we will call ‘identified elements’ the results obtained by the
users at the given step. For example, the identified elements are functions for the first
step, technical problems for the second step, requests for the third step, etc.
5.1.1 For the total population
The number of identified elements appears to be following a two-step tendency (Table
3).
Table 3. Paired-mean comparison of the number of elements given by the total population
Elements identified at each
step
Total (N=42)
mean (stdev)
mean of
comparison
p-value (t-test)
Biomimetic
tool
1. Functions
5.14 (1.86)
-
-
None
2. Technical problems
6.90 (2.79)
+1.76
1.10-08***
None
3. Requests
9.69 (5.18)
+2.78
3.10-05***
Taxonomy
4. Organisms
7.35 (3.05)
-2.33
2.10-03**
AskNature
5. Biological strategies
6.95 (2.92)
-0.40
8.10-03**
AskNature
6. Technical solutions
5.28 (2.24)
-1.66
4.10-05***
AskNature
P-value: ‘***’ < 0.001; ‘**’ < 0.01; ‘*’ < 0.05; ‘.’ < 0.1
From the problem analysis (step 1) to the transposition to biology (step 3) we observe a
significant increase in the number of elements given by the users whereas it significantly
decreases from the identification of organisms (step 4) to the transposition to technical
solution (step 6). This first pattern leads us to consider the fourth step as a bottleneck for
the biomimetic process used in this experiment.
In order to get more information on this pattern we looked at the link between the
different steps through a correlation analysis (supplementary table 1). The results of the
correlation analysis appear coherent with a two-part pattern, as they show strong
correlations between the first three steps on one side (corr. Min. is for step 1 and 3,
correlation = 0.512, Spearman test, p-value = 5.10-04) and between the last three steps on
the other side (corr. Min. is for step 4 and 6, correlation = 0.664, Spearman test, p-value
= 2.10-06). Outside those two clusters, some correlations appear significant but they are
less strong (correlations below 0.5). These tendencies support the idea that the fourth step
constitutes a bottleneck for the participants at the overall scale. Interestingly, the shape of
this pattern is reminiscent of the diamond-shape process of creativity, consisting of a
divergent phase, where every angle of an issue is considered, before leading to a
convergent phase, where the final solution is determined, within the pool of previously
identified strategies.
5.1.2 Depending on the users’ background
If we look at the same elements partitioned by backgrounds, we can see that the number
of elements identified is significantly higher for biologists at each step, except at the third
step (Table 4).
Table 4. Comparison of the number of elements given at each step of the process
partitioned by profile
Elements identified at each
step
Biologists (n=21)
mean (stdev)
Engineers (n=21)
mean (stdev)
p-value
(t-test)
Biomimetic
tool
1. Functions
5.85 (2.00)
4.42 (1.43)
1.10-02*
None
2. Technical problems
8.14 (3.16)
5.66 (1.65)
1.10-03**
None
3. Requests
10.85 (5.66)
8.52 (4.47)
1.10-01
Taxonomy
4. Organisms
8.66 (3.26)
6.04 (2.22)
4.10-03**
AskNature
5. Biological strategies
8.33 (2.93)
5.57 (2.22)
1.10-03**
AskNature
6. Technical solutions
6.42 (1.43)
4.14 (2.35)
5.10-04***
AskNature
Bold: Highest values, P-value: ‘***’ < 0.001; ‘**’ < 0.01; ‘*’ < 0.05; ‘.’ < 0.1
We then wondered about the mechanisms behind these differences, are they due to an
initial higher number of elements identified by biologists that are carried over to the
following steps? Or do they appear at each step? Only at specific steps?
Firstly, it appears that differences between both populations aren’t significant on
the number of requests formulated during the third step. However, results are significantly
different during the identification and selection of potential biological models of interest,
leading us to think that biologists are more efficient during this fourth step. To confirm
this result, we performed a correlation analysis on biologists’ results (supplementary table
2). The same two distinct phases above-mentioned (part 5.1.1. diamond shaped tendency)
can be identified in the results (supplementary table 2) and the step 3 and 4 aren’t
significantly correlated (corr.: 0.407, Spearman test, p-value = 7.10-02), confirming our
hypothesis on the presence of at least two steps, the problem analysis and the
identification and selection of potential biological models of interest, of higher efficiency
linked with a background in biology.
We also performed a statistical analysis to compare the evolution of the number
of elements between steps for both populations. All the differences where significant for
users with a background in engineering, meaning that, as it was identified at the general
level, the selection of organisms on AskNature (step 4) appears to be the first bottleneck,
as the first step of significant decrease, for this population.
However, when we look at the same analysis for users with a background in
biology (Table 5), we can see that the differences between steps 3 and 4, and at a higher
level, between steps 4 and 5 are not significant.
Table 5. Paired-mean comparison of the number of elements given by biologists
Elements identified at each
step
Biologists (n=21)
mean (stdev)
mean of
differences
p-value (t-test)
Biomimetic
tool
1. Functions
5.85 (2.00)
-
-
None
2. Technical problems
8.14 (3.16)
+2.28
4.10-06***
None
3. Requests
10.85 (5.66)
+2.71
5.10-03**
Taxonomy
4. Organisms
8.66 (3.26)
-2.19
6.10-02
AskNature
5. Biological strategies
8.33 (2.93)
-0.33
8.10-02
AskNature
6. Technical solutions
6.42 (1.43)
-1.90
3.10-03**
AskNature
P-value: ‘***’ < 0.001; ‘**’ < 0.01; ‘*’ < 0.05; ‘.’ < 0.1
It appears that a training in biology can lead to a better handling of the steps which are
responsible for the decrease of the number of elements, i.e. the bottlenecks of the process.
Thus, under these circumstances, the previously identified pattern appears to be modified.
It is especially interesting to see that differences between the number of strategies
identified from biological organisms isn’t significant for biologists where it is for
engineers (-0.33 vs -0.45, t-test, p-value = 8.10-02 vs 4.10-02) confirming the results
underlined previously. Biologists almost always transfer a biological strategy from an
identified biological organism.
This overall analysis has led us to consider some quantitative differences
depending on the background of the users. More especially, it showed that biologists are
better equipped to face step 4 and 5 of the biomimetic process.
5.2 Step by step analysis
After this first analysis, we will examine the process in more depth in order to search for
cognitive explanation leading to the results previously presented.
5.2.1 Step 1. Problem analysis
The 1st step of the experimental process aims at describing the problem through the
identification of the constraints and the consumers’ real needs in order to formalise
technical system’s functions.
During this first step, users gave a significantly different number of functions
depending on the user’s background (Table 4). As a result, we tried to characterise this
difference by looking at the distribution of the identified functions for each background.
First of all, in total, the population with a background in biology identified 17
functions where the population with a background in engineering identified 14 functions.
Moreover, even if the results don’t appear significantly different, 13 functions have a
higher percentage of identification by biologists (supplementary table 3). Thus, the largest
number of functions identified by biologists overall is due to the accumulation of small
and disseminated differences.
In conclusion, the data suggests that biologists show a greater diversity of
identified functions, but don’t present a strong tendency toward specific ones.
Through the analysis of each function, users had to determine a set of Technical
Problems (TP) which is studied in the next section.
5.2.2 Step 2. Abstraction of technical problems
The 2nd step of the experimental process aims at breaking down the identified functions
into their several technical problems. For example, some users decomposed ‘thermic
insulation’ into ‘the insulation from solar radiation’, ‘heat conduction’, ‘heat
convection’, etc.
For each function, except noise insulation, users gave a non-significantly different
number of TP (supplementary table 4). We then wondered if some TP were identified
specifically by one background or the other. We observed that 61.54% of them have been
identified by both populations, 30.77% identified only by users with a background in
biology and 7.69% only by users with a background in engineering (Chi2 test, p-value =
0.035).
Even if, as previously explained, the distribution doesn’t appear significantly
different overall, the TP specifically identified by biologists are divided between 10
functions, where those specifically identified by engineers represent only four functions.
This leads us to think that biologists have a wider range of consideration while analysing
a technical problem. The increase of diversity reduces the number of identifications for
each function, and so their statistical ‘weight’, which can partly explain why the results
aren’t significant.
The identified TP are then formalised thanks to the biomimicry taxonomy into
request during the third step.
5.2.3 Step 3. Transpose to biology
The 3rd step of the experimental process aims at transposing technical problems into
biological requests through the use of the biomimicry taxonomy. For example, ‘insulation
against solar radiation’ have been transposed into several requests like: ‘protect from
non-living threats (light)’, ‘modify light/colour’, ‘transform radiant energy’, etc.
The results observed on the formulation of requests are similar than those on TP,
the statistical analysis shows no significant differences on the overall functional
distribution of requests. Again, if we characterise the requests, we can see that 60% have
been identified by both populations, 33.3% identified only by biologists and 6.6%
identified only by engineers (Chi2 test, p-value = 0.027). Besides, it appears for both
populations that the functions targeted by the TP and the requests are similar.
Moreover, when we considered the number of identified requests by TP for each
function, the comparison led us to only one function with a significant difference: the
biotic insulation (1.5 vs 1, t-test, p-value = 7.10-3). At first sight, the insulation against
living beings might not appear central here, but it is actually crucial in Ivory Coast and
other countries where animals and especially insects like mosquitoes transmit diseases
like yellow fever, dengue or malaria.
The increase in diversity induced by a background in biology and observed on the
first three steps of the process is a strong argument supporting the potential for innovation
brought by biologists on those engineering-centred steps.
The second part of the experiment focuses on the identification, selection and
understanding of biological strategies.
5.2.4 Step 4. Identification and selection of biological model(s) of interest
The 4th step of the experimental process aims at identifying and selecting biological
organisms by entering the requests formalised during the 3rd step into AskNature’s search
engine. For example, the request ‘protect from non-living threats (light)’ led to the
identification and selection of the Sphicterochila boisseri a specie of desert snail.
First of all, at a general scale only 12 of the 17 initial functions led to at least one
organism. Among these functions, functions 1 and 2 represent 74.85% of the selected
organisms. We thought of two hypotheses explaining this tight distribution. Either the
results are linked with the users’ choices: facing the short time frame of the experiment,
they focused on the two main functions. Or, the results are linked with AskNature which
doesn’t display an equivalent number of, or accessibility to, organisms for each function.
As pointed out in the table 3, biologists gave a significantly higher mean number
of organisms (8.66 vs 6.04, t-test, p-value = 4.10-03**). In order to point out which
parameters led to an identification by a particular population, we characterised the
organisms’ strategies. The characterisation was made on the type of associated strategy
(Material/Process, Shape, System), the scientific field (Physics, Chemistry, Biology), the
scale (Macro, Meso, Micro), the type of kinetic (Dynamic, Static), the energy
consumption (Active, Passive) and the type of organisms (Bacteria, Fungus, Vegetal,
Insects, Other animals). Table 6 displays the parameters showing significant results on
the characterisation partitioned by profile.
Table 6. Comparison of the characteristics of all organisms’ associated strategies
partitioned by profile (n = 21)
Material
Process
Shape
Physics
Chemistry
Micro
Static
Passive
Vegetal
mean (stdev)
Biologists
6.90
(2.77)
2.67
(1.31)
6.85
(2.26)
1.85
(1.35)
5.67
(2.49)
7.80
(2.58)
8.90
(3.08)
4.14
(1.68)
Engineers
4.71
(1.84)
1.67
(1.15)
4.90
(2.07)
1.00
(1.14)
3.61
(1.71)
5.14
(2.65)
5.61
(2.59)
2.28
(1.84)
P-value
(t-test)
4.10-03**
1.10-02*
5.10-03**
3.10-02*
3.10-03**
2.10-03**
5.10-04***
1.10-03**
Bold: Highest values, P-value: ‘***’ < 0.001; ‘**’ < 0.01; ‘*’ < 0.05; ‘.’ < 0.1
Thus, we can see that there is a significantly higher number of organisms identified by
biologists for some categories of each studied parameter, underlining their higher
efficiency during the fourth step.
On the other hand, when we consider the same analysis but with proportions, there
are no significant differences between the distribution of both populations suggesting that
we cannot underline, based on this dataset, background-related patterns about the
cognitive choices of organisms.
A first hypothesis is that the identified organisms are always the same because of
their indexation, through the AskNature search engine, on the first pages of the database,
leading to a smoothing of the results because of a strong redundancy.
A second hypothesis is that the selection process is driven by a parameter
independent from those used in the analysis. As a result, a higher efficiency, or a lengthier
time of experiment, would lead to a higher number of identified organisms but to the same
proportional distribution.
So far, these results lead us to think that biologists are more efficient during the
fourth step, but no specific cognitive tendencies have been identified. In order to prevent
the smoothing effect caused by the redundancy of organisms identified by both
populations, we determined two categories of organisms, those specifically identified by
only one profile and those identified by at least one user of both populations (Table 7).
Table 7. Comparison of the number of organisms’ associated strategies for each category
partitioned by profile (n = 21)
Mean (stdev)
Mean specific (stdev)
Mean mutual (stdev)
Biologists
9.71 (3.40)
4.00 (2.88)
5.71 (2.00)
Engineers
6.71 (2.72)
2.28 (1.95)
4.42 (2.13)
P-value (t-test)
3.10-03**
3.10-02*
5.10-02*
Bold: Highest values, P-value: ‘***’ < 0.001; ‘**’ < 0.01; ‘*’ < 0.05; ‘.’ < 0.1
Biologists gave a higher number of elements for both categories but again, when analysed
in proportion, these differences don’t appear significant. Thus, users from each
background have on average the same percentage of organisms mutually and specifically
identified.
In order to understand the mechanism under the specific choices made by both
populations, we focused on the background-specific organisms. Overall, it appears that
specific and mutual organisms are differentiated depending on the way they address the
different functions. Figure 4 describes the organisms’ partition between the four most
identified functions and a cluster composed of all the other secondary functions.
Figure 4. Comparison of organisms’ functional distribution depending on ‘mutual’ (A) or
‘specific’ (B) categories
Where the distribution of the organisms identified in common appears tight, centred on
F1 and F2, the distribution of organisms identified specifically appears significantly
different (Chi2 test, p-value =1.10-04***) more diverse, targeting some of the secondary
functions along with the main ones.
This difference of partition can also be seen on the proportion of mutually and
specifically identified organisms for each function (ex for F4: 81.82% are specific, Chi2
test, p-value = 3.10-05***) (supplementary table 5).
We then performed an analysis on the characteristics previously presented in order
to obtain more information on the cognitive aspects of the choice of specific organisms
(Figure 5).
Figure 5. Comparison of ‘specific’ or ‘mutual’ organisms’ associated strategies, p-
value: ‘***’ < 0.001; ‘**’ < 0.01; ‘*’ < 0.05; ‘.’ < 0.1
Through this analysis, we were able to characterise the main differences between the
organisms identified by both schools and those identified by only one or the other.
The next step is to compare within the specific category, the choices made by
biologists and engineers. Similar to the previous analysis, if we look at the quantitative
aspect, biologists and engineers gave strongly different results, but those results don’t
appear significantly different in proportion. The only two parameters that show
significant differences in proportion are the scale (microscopic, 71% vs 56%, Chi2 test,
p-value = 3.10-2) and the energy consumption (passive, 90% vs 75%, Chi2 test, p-value =
3.10-3), meaning that biologists drive the search for solutions toward the microscopic
world and passive strategies.
In terms of cognitive pathway, the microscopic world ultimately calls for specific
knowledge where macroscopic organisms can be more easily apprehended by non-
biologists, potentially explaining these results. It also shows that even within this
restrictive and determined frame of research (AskNature), there is a set of organisms
which identification appears background-related.
Once identified and selected, the strategies of living beings have to be analysed
and understood in order to be subsequently transferred into technological innovation. As
explained in part 4.2, we compared the level of understanding of these strategies by
focusing on the not-understood organisms.
5.2.5 Step 5. Abstraction of biological strategies
The 5th step of the experimental process aims at deeply analysing AskNature’s data in
order to abstract the selected biological strategies. Depending on the level of
understanding of the analysed strategies, the precision of the modelization varies. For
example, the S. boisseri (desert snail) protects himself from light through a highly
reflective shell. A first level of understanding would be to focus on the overall idea and
so abstract the strategy by considering a protection through a highly reflective panel, but
it doesn’t give actual information on the reflective properties of the snail shell. A deeper
level of analysis, on colours, nanostructures, shapes, etc. can lead to a more informative
model and so to a more comprehensive abstraction step of the underlying biological
strategy, making this 5th step a critical step of the biomimetic process.
The proportion of not understood strategies appeared significantly different
depending on the background of the users (0.6% vs 12.3%, Chi2 test, p-value = 5.10-5).
As we can expected, the absence of a background in biology leads to a higher rate of not-
understood biological strategies.
We then wondered about the distribution of these strategies. More precisely, we
wanted to know if they were linked with the ‘specific’ or ‘mutual’ categories above
mentioned. The results show that 64.29% of the errors (nine organisms) were made by
engineers on ‘specific’ strategies, where no mistakes have been made by biologists on
‘specific’ strategies. These errors represent 18.75% of the total number of organisms
specifically identified by engineers. As a result, biological data made available by
biologists appear more relevant and reliable for the generation of abstraction model. This
last result underlines that a lack of background in biology can drive the resolution process
toward irrelevant or misunderstood biological strategies, resulting in the loss of resources
and potentially in the failure of the innovative process implementation.
Altogether, the results of the fifth step point out the asset that represents a
background in biology for the relevance and reliability of the chosen data on which is
based the abstraction process of the biological strategies.
Through this step by step analysis, we pointed out the potential contribution of biologists
during the biomimetic process. As previously explained, the step 6 isn’t detailed in this
step by step analysis. Because of the short time frame of the experiment results obtained
in the step 6 were used as markers of understanding but they weren’t precise enough to
be separately analysed.
This article deals with the biologists’ contribution during the practice of
biomimetics in order to foster his integration during both the practice of biomimetics and
the design of the method and tools. As a result, we also investigate the users’ feedbacks
toward the tools used during the experiment in order to underline any shortcomings.
5.3 Tool analysis and users’ feedbacks
The tools' evaluation was performed by the users with a 5-point Likert scale on
satisfaction, learnability, error, wealth of information and precision of information.
Firstly, we evaluated the Biomimicry Taxonomy (Table 8).
Table 8. Biomimicry Taxonomy’s evaluation by both profiles (n = 21)
Learnability
Errors
Wealth of information
Precision
Satisfaction
median (biologists)
4
4
3
3
4
median (engineers)
4
4
4
3
4
p-value (MWW test)
6.10-1
4.10-1
2.10-3**
8.10-1
3.10-1
Bold: Highest values, P-value: ‘***’ < 0.001; ‘**’ < 0.01; ‘*’ < 0.05; ‘.’ < 0.1
As we can see in table 8, the Biomimicry Taxonomy obtained overall positive grades.
The precision and wealth of information are the criteria which obtained the lowest grade
for both populations underlining some of the feedbacks given by the users: ‘The taxonomy
limits the reflection’, ‘Hard to find matching requests for some of the TP’. We can see
that the only criterion which obtained a significantly different grade is the wealth of
information (MMW test, p-value = 2.10-3**). These results lead us to think that even if
engineers acknowledged the diversity of the requests available through the taxonomy,
they didn’t find the tool very precise and had troubles formulating requests addressing
the identified TP directly. On the other hand, biologists felt restricted by the tool in both
its diversity and precision.
We then evaluated AskNature following the same protocol (table 9).
Table 9. AskNature’s evaluation by both profiles (n = 21)
Learnability
Errors
Wealth of information
Precision
Satisfaction
median (biologists)
5
4
4
3
4
median (engineers)
5
4
4
4
4
p-value (MWW test)
9.10-2
7.10-1
1.10-1
2.10-3**
7.10-1
Bold: Highest values, P-value: ‘***’ < 0.001; ‘**’ < 0.01; ‘*’ < 0.05; ‘.’ < 0.1
AskNature's ease of access has been underlined by all users. Interestingly, engineers gave
high scores on all criteria during AskNature’s evaluation, with a significant difference
comparing to biologists on its precision (MWW test, p-value = 2.10-3**) (Figure 6).
Figure 6. Comparison between scores given by biologists (A) and engineers (B) on
AskNature’s precision
On one hand, engineers say the tool is rich, precise and satisfying. On the other hand,
biologists show a mixed opinion on the wealth of information and the precision of the
tool. Two main hypotheses can explain such results.
Firstly, because of the differences on their respective knowledge in biology, both
populations don’t have the same expectations for such tool. As the database does not list
all known organisms and does not provide enough information to understand all the
phenomena, biologists felt frustrated ‘the search engine restrains imagination’, ‘the
website limits, frames and offers very little ideas’ and gave lower scores during its
evaluation. Where biologists are already aware of the diversity and complexity of
biological strategies, engineers might feel overwhelmed by data they aren’t accustomed
to deal with. Hence, this feeling can lead to a Wahoo effect driving engineers to over-rate
the tool. As a result, the lack of a background in biology leads to a variation in the frame
of reference which is a break in the innovative process because it will narrow the spectrum
of research and hence the potential for novelty.
A second aspect is that engineers and biologists may not have the same goal while
using the data. The reasoning processes of biologists and engineers intrinsically differ in
their objectives: where the main objective of biologists is to discover and understand how
entities work, the main objective of engineers is to innovate and design new entities.
These fundamentally different aims influence their reasoning, habits and insights when
facing data. During the biomimetic design process, engineers may consider biological
phenomena as a source of inspiration to be used during creativity steps rather than a
source from which they can directly extract technical innovations. Thus, the expectations
on the depth of understanding, and, therefore, the diversity and precision of the data, can
vary, which might explain the results.
6. Comparison with experts’ opinion
As presented in section 4, the results that are discussed in this section focus on only two
axes of the survey. The first question was about the ‘necessity’ of biologists during the
practice of biomimetics. As the term necessity appears quite restrictive, we expected a
mitigate answers by the experts. Three answers were available: ‘Yes’, ‘No’, ‘Depends on
the project’. 53% of the respondents indicate that the necessity of a biologist during the
practice of biomimetics depends on the project, 41% answered ‘Yes’ and 6% answered
‘No’. It is essential to underline that they nonetheless all recognise the asset that
represents a biologist during the practice of biomimetics. When asked why biologists
aren’t necessary during the biomimetic design process, experts point out three main
aspects. First, the available biological data might be detailed and clear enough to be
understood by non-biologists. Secondly, some design projects don’t require a deep
understanding of the biological strategy to successfully meet their goals. And finally, the
biological expertise doesn’t have to be necessarily provided by a biologist but can be a
skill own by another profile.
When we consider a parallel between the results of our case study and the answer
to the survey, we can see that in both situations, an expertise in biology emerge as an
asset for the practice of biomimetics and the understanding of biological data. Moreover,
the existence of several levels of inspiration depending on the background, is also
supported by both the results from the case study and the survey. Finally, the question of
the holder of the biological knowledge emerges as fundamental and will be detailed in
the discussion part.
The second question is about the role of biologists during the biomimetic process.
The results from the case study allowed us to identify three main contributions: the ability
to provide a deep understanding of biological data, to make reliable choices, and to offer
diversity during the engineering centred steps thanks to an alternative way of reasoning.
In order to evaluate the relevance of our findings, we proposed five different answers
(and an ‘other’ response) and restricted to a maximum of three answers by the respondent
(Figure 7).
Figure 7. Experts’ opinion on the contribution of biologists during biomimetic process
These results support two of our findings as they both the understanding and the
identification show more than 80% of the selection. Interestingly, no experts identified
missing aspects. Specifically, the biologist’s ability to add variability isn’t identified if
not proposed. Taken together with our first results, this aspect suggest that the added
variability observed during the case study is presumably not specific to biologists. It is
indeed a well-known fact in design that pluridisciplinarity in general helps overcome
fixations and brings diversity. However, biologists do bring variability and as a result it
can be considered as an added-value, even if non-specific. Experts also pointed out the
biologists’ role as bridges between biology and engineering through the translation of
biological vocabulary and concepts. This aspect hasn’t been studied in the case study as
the experiment was performed individually.
Altogether, the experts’ opinion, strengthen our analysis on the asset that
represents an expertise in biology for the biomimetics process. The next part will discuss
our findings, their relevance, their consequences as well as our future axes of research.
7. Discussion
This article presented the contribution of biologists during the biomimetic design
process through a case study and a survey of experts’ opinion. Our results underlined why
biologists represents assets and so levers for optimisation for the biomimetic design
process.
Nevertheless, several limitations have to be pointed out. First, the case study has
been performed by the individual student and not interdisciplinary teams. As a result,
both positive or negative aspects of interdisciplinary work, like constructive interactions
or communication challenges, aren’t taking into account in this study.
Furthermore, this article doesn’t tackle one other challenging step of the
biomimetic process which concerns the means of identification of biological organisms.
Indeed, only AskNature was used as the source of organisms but as previously underlined,
even if it represents the greatest biomimetic database available, AskNature is very
restrictive. Without AskNature, is an engineer able to find organisms of interest? The
strong concurrence in the industrial context will lead biomimetic teams to consider much
more organisms than those listed on AskNature, pointing out another strong need for
biological expertise.
The short time frame of the study and the fact that, even if supported by experts’
answers, the sample is composed of students and not professionals have to be taken into
consideration. In that respect, we are currently working on a long-term project with
industrial partners in order to confirm our findings, test new hypotheses and
methodological innovations.
Finally, the design problem chosen in this article is easily accessible because not
very technical. Therefore, it has to be considered that these results are linked with the
studied subject. Notably, the steps 1 to 3 are well managed by biologists and steps 4 to 6
are well managed by engineers, all partially due to the topic of the case study which was
easily accessible. We are aware of those biases but, since it makes technical issues more
accessible for both populations, an easy design problem might reduce the gaps between
both populations rather than exaggerate them, and so doesn’t represent a problematic
factor in this context.
Our findings open several research axes. Where this article aims at foregoing the
integration of biological expertise, it doesn’t show that biologists can replace engineers.
Each profile brings its own specificities and we believe that it is the cooperation between
both profiles that will lead to a successful implementation of biomimetics.
The terms and form of such a cooperation are to be entirely defined. As pointed
out by the experts: should we actually consider the integration of ‘biologists’ or of
‘biological expertise’ and if we are, which ‘biologists’ are we talking about? A
transdisciplinary profile, having both a biological expertise, focused on the needs of
biomimetics design, and knowledge in engineering in order to link both scientific fields
appears as a challenging but rewarding research goal. It also raises the question of the
methodological framework surrounding the integration of such a profile. As our findings
suggest, an expertise in biology should facilitate the access to a deeper level of analysis
but it also requires tools and processes adapted for such approaches and pluridisciplinary
work.
Conclusion
Biomimetics isn’t a new strategy but its implementation remains highly limited in the
industrial sector. Our first assessment is that industrial companies that want to use
biomimetics as an innovation strategy need guidance. The complexity and
transdisciplinary aspect of this approach cannot be ignored or considered secondary.
Biomimetics’ complexity and novelty potentially require new profiles, structural
evolutions and specific training. More specifically, our observation is that the current
methodological framework is based on biomimetic tools without taking into account
biologists’ expertise, leading to processes that aren’t made to be transdisciplinary in
practice.
However, we also pointed out that numerous researchers underlined the necessity
of biologists during biomimetic processes. This contradiction raises the research question
tackled in this article: Facing the current biomimetics’ methodological framework, does
knowledge in biology represent an asset for the practice of biomimetics? Our hypothesis
is that it does. The results presented in this article underlined two main aspects. Firstly,
biologists appear to be more effective during the use of biomimetic tools as they’ve found
more elements and have made less mistakes. Secondly, both biologists and engineers
bring their way of reasoning leading to an increase in the diversity of identified biological
models. These results are supported by experts’ opinion gathered through a survey and
so validate our hypothesis. By showing, in an accessible and restricted context, that
biologists expertise contributes to the innovative nature of biomimetics, these findings
contribute to the consideration, by engineering designers, of biologists’ role during the
design of a transdisciplinary methodological framework for the practice of biomimetics.
Choosing between the development of biomimetic tools and the integration of biological
expertise has proven to be a mistake as biomimetics is still struggling to spread in the
industrial world. We advocate that both aspects need to be investigated and combined in
order to solve the current paradoxical situation and take biomimetics to the next level.
This article is built as the foundation stone for our future work on biomimetic
design teams, process and tools, combining expertise in engineering and biology. In this
perspective, our coming work focuses on collecting information about engineers’ and
biologists’ needs during the biomimetic process in order to identify key elements
obstructing the integration of a profile with a background in natural sciences. We will
then work on providing a more precise methodological support, discussing the biological
knowledge of interest in order to target which profile to integrate in biomimetic teams,
when and for what purpose should they be integrated, and using which tools. The overall
aim of our work is to optimise the methodological framework within a transdisciplinary
context in order to guide industrial actors who want to implement biomimetics as an
innovation strategy.
Acknowledgments. We thank the students who took part in our experiment and the respondents
to our questionnaire. Special thanks to Juliette Gambaretti for her guidance and support on
statistical analyses.
Disclosure statement
No potential conflict of interest was reported by the author.
References
Ahmed-Kristensen, S., B.T. Christensen, and T.A. Lenau. 2014. “Naturally Original:
Stimulating Creative Design through Biological Analogies and Random
Images.” In Design Conference, 427–36. Dubrovnik.
Altshuller, G.S. 1984. Creativity as an Exact Science: The Theory of the Solution of
Inventive Problems. Gordon & Breach Science Publishing, New York. Gordon
and Breach Science Publishers.
Bar-Cohen, Y. 2006. “Biomimetics - Using Nature to Inspire Human Innovation.”
Bioinspiration and Biomimetics 1 (1): P1–12. doi:10.1088/1748-3182/1/1/P01.
Bar-Cohen, Y. 2012. “Biologically Inspired Technologies for Aeronautics.” In
Innovation in Aeronautics, 15–36. Woodhead Publishing.
doi:10.1533/9780857096098.1.15.
Baumeister, D. 2014. Biomimicry Resource Handbook: A seed bank of knowledge and
best practices. 2014th ed. CreateSpace Independent Publishing Platform.
Benyus, J.M. 1997. Biomimicry: Innovation Inspired by Nature. Quill.
Bila-Deroussy, P. 2015. “Systemic approach of creativity : Tools and methods to
address complexity in design.” ENSAM - Paris.
Biomimicry Institute. 2002. “AskNature - Innovation Inspired by Nature.” AskNature.
https://asknature.org/.
Boeuf, G. 2007. “Océan et recherche biomédicale.” Journal de la Société de Biologie
201 (1): 5–12. doi:10.1051/jbio:2007001.
Bourgeois, P. 2007. Les grands défis technologiques et scientifiques au XXIe siècle.
Ellipses.
Chakrabarti, A., P. Sarkar, B. Leelavathamma, and B.S. Nataraju. 2006. “A Functional
Representation for Aiding Biomimetic and Artificial Inspiration of New Ideas.”
Artificial Intelligence for Engineering Design, Analysis and Manufacturing:
AIEDAM 19 (2): 113–32. doi:10.1017/S0890060405050109.
Chirazi, J., K. Wanieck, P.E. Fayemi, C. Zollfrank, and S. Jacobs. 2019. “What Do We
Learn from Good Practices of Biologically Inspired Design in Innovation?”
Applied Sciences 9 (4). Multidisciplinary Digital Publishing Institute: 650.
doi:10.3390/app9040650.
Deldin, J. M., and M. Schuknecht. 2013. “The AskNature Database: Enabling Solutions
in Biomimetic Design.” In Biologically Inspired Design, 17–28. Springer,
London. doi:10.1007/978-1-4471-5248-4_2.
Eggermont, H. 2015. “Nature-Based Solutions: New Influence for Environmental
Management and Research in Europe.” Gaia 24 (4): 243–48.
doi:10.14512/gaia.24.4.9.
Faludi, J. 2017. “Golden Tools in Green Design: What Drives Sustainability,
Innovation, and Value in Green Design Methods?” University of California
Berkeley. https://digitalcommons.dartmouth.edu/facoa/2784.
Fayemi, P.E. 2016. “Innovation through Bio-Inspired Design : Suggestion of a
Structuring Model for Biomimetic Process and Methods.” ENSAM - Paris.
Fayemi, P.E., N. Maranzana, A. Aoussat, and G. Bersano. 2014. “Bio-Inspired Design
Characterisation and Its Links with Problem Solving Tools.” In Design
Conference, 173–82. Dubrovnik.
Goel, A.K., S. Vattam, B. Wiltgen, and M.E. Helms. 2014. “Information-Processing
Theories of Biologically Inspired Design.” In Biologically Inspired Design,
127–52. Springer, London. doi:10.1007/978-1-4471-5248-4_6.
Gordon, W. J. J. 1961. Synectics - The Developmnent of Creative Capacity. New York;
Evanston: Harper & Row.
Grabowski, H., R. S. Lossack, and E. F. El-Mejbri. 1999. “Towards a Universal Design
Theory.” In Integration of Process Knowledge into Design Support Systems, 49–
56. Dordrecht: Springer, Dordrecht. doi:10.1007/978-94-017-1901-8_2.
Guerrero, J. E., D. Maestro, and A. Bottaro. 2012. “Biomimetic Spiroid Winglets for
Lift and Drag Control.” Comptes Rendus - Mecanique 340 (1–2): 67–80.
doi:10.1016/j.crme.2011.11.007.
Helfman Cohen, Y., and Y. Reich. 2016. Biomimetic Design Method for Innovation and
Sustainability. Biomimetic Design Method for Innovation and Sustainability.
doi:10.1007/978-3-319-33997-9.
Hoagland, M.B., and B. Dodson. 1995. The Way Life Works. 1st ed. New York: Crown.
Hwang, J., Y. Jeong, J. W. Hong, and J. Choi. 2015. “Biomimetics: Forecasting the
Future of Science, Engineering, and Medicine.” International Journal of
Nanomedicine 10: 5701–13. doi:10.2147/IJN.S83642.
IPCC. 2014. “Climate Change 2014: Synthesis Report. Contribution of Working
Groups I, II and III to the Fifth Assessment Report of the Intergovernmental
Panel on Climate Change on Climate Change.” IPCC, Geneva, Switzerland.
ISO/TC266. 2015. “Biomimétique -Terminologie, Concepts et Méthodologie.”
https://www.iso.org/fr/committee/652577/x/catalogue/.
Keshwani, S., T.A. Lenau, S. Ahmed-Kristensen, and A. Chakrabarti. 2017.
“Comparing Novelty of Designs from Biological-Inspiration with Those from
Brainstorming.” Journal of Engineering Design 28 (10–12): 654–80.
doi:10.1080/09544828.2017.1393504.
Lahonde, N. 2010. “Design Process Improvement : Proposal of a Model for Design
Methods Selection to Support the Decision.” ENSAM - Paris.
Lee, J., J. Pries-Heje, and R. Baskerville. 2011. “Theorizing in Design Science
Research.” In Proceedings of the 6th International Conference on Service-
Oriented Perspectives in Design Science Research, 1–16. Springer-Verlag.
doi:10.1007/978-3-642-20633-7_1.
Lenau, Torben Anker. 2009. “Biomimetics as a Design Methodology - Possibilities and
Challenges.” International Conference on Engineering Design, ICED, 121–32.
Lindemann, U., and J. Gramann. 2004. “Engineering Design Using Biological
Principles.” In Design Conference, 355–60. Dubrovnik.
Magyar, A., V.K.A. Arthanareeswaran, L. Soós, K. Nagy, A. Dobák, I.M. Szilágyi, N.
Justh, A.R. Chandra, B. Köves, and P. Tenke. 2017. “Does Micropattern
(Sharklet) on Urinary Catheter Surface Reduce Urinary Tract Infections? Results
from Phase I Randomized Open Label Interventional Trial.” European Urology
Supplements 16 (3): e146–48. doi:10.1016/S1569-9056(17)30153-7.
Nagel, J.K.S., R.L. Nagel, R.B. Stone, and D.A. McAdams. 2010. “Function-Based,
Biologically Inspired Concept Generation.” Artificial Intelligence for
Engineering Design, Analysis and Manufacturing: AIEDAM 24 (4): 521–35.
doi:10.1017/S0890060410000375.
Nielsen, J. 1993. Usability Engineering. Academic Press.
Ohno, T. 1978. Toyota Production System: Beyond Large-Scale Production.
Productivity Press. 1 edition. Vol. 1. Cambridge, Mass: Productivity Press.
Richter, I. A., T. Wells, and M. Kemp. 2008. Notebooks. Oxford World’s Classics. New
Editio. Oxford, New York: Oxford University Press.
Sachs, J. D. 2015. The Age of Sustainable Development. Columbia University Press.
doi:10.7312/sach17314.
Schöfer, M. 2015. “Processes and Methods for Interdisciplinary Problem Solving and
Technology Integration in Knowledge-Intensive Domains.” ENSAM - Paris.
Schöfer, M., N. Maranzana, A. Aoussat, G. Bersano, and S. Buisine. 2018. “Distinct
and Combined Effects of Disciplinary Composition and Methodological Support
on Problem Solving in Groups.” Creativity and Innovation Management 27 (1):
102–15. doi:10.1111/caim.12258.
Snell-Rood, E. 2016. “Interdisciplinarity: Bring Biologists into Biomimetics.” Nature
529 (7586): 277–78. doi:10.1038/529277a.
Tinsley, A, P Midha, R.L. Nagel, and D.A. McAdams. 2007. “Exploring the Use of
Functional Models As a Foundation for Biomimetic Conceptual Design.” ASME
2007 International Design Engineering Technical Conferences and Computers
and Information in Engineering Conference, 1–15. doi:10.1115/DETC2007-
35604.
Vandevenne, Dennis, Paul Armand Verhaegen, Simon Dewulf, and Joost R. Duflou.
2014. “A Scalable Approach for Ideation in Biologically Inspired Design.”
Artificial Intelligence for Engineering Design, Analysis and Manufacturing:
AIEDAM 29 (1): 19–31. doi:10.1017/S0890060414000122.
Vincent, J.F.V., O.A. Bogatyreva, N.R. Bogatyrev, A. Bowyer, and A.K. Pahl. 2006.
“Biomimetics: Its Practice and Theory.” Journal of the Royal Society Interface 3
(9): 471–82. doi:10.1098/rsif.2006.0127.
Wanieck, K., P.E. Fayemi, N. Maranzana, C. Zollfrank, and S.R. Jacobs. 2017.
“Biomimetics and Its Tools.” Bioinspired, Biomimetic and Nanobiomaterials 6
(2): 53–66. doi:10.1680/jbibn.16.00010.
Zhang, G. 2012. “Biomimicry in Biomedical Research.” Organogenesis 8 (4): 101–2.
doi:10.4161/org.23395.
