BioInspired! So What? Now What!

Although lack of information about biological systems is often considered a key inhibitor, tools and repeatable methods are also essential if we are to become proficient at bio-inspired design.  When new fields are developed, existing methodologies can sometimes be adapted from other disciplines that are comparable or familiar to designers. A better approach involves studying how designers go about designing within the new field. 

As described in September 2010 newsletter, “Research on design cognition is interested … in how … designers behave, think, perceive, act, communicate and learn when designing”.  It investigates what designers bring to a situation, the models they use and develop, how their work relates to the world around them and how they collaborate.  In comparison to a generalized ‘best practices’ approach, design cognition takes into account the diversity of designers and the context in which they work. 

There are other fields (such as design theory) that explore similar issues. Research on design cognition makes specific commitments both about the research end products and about its methods of investigation. The products of a design cognition study should include a detailed and consistent account of the information processing occurring in the design phenomenon or behavior. The methods should be empirical and evidence-based, ranging from observational studies to controlled experiments, neuro-imaging and computer simulations.

For four years, the Georgia Tech’s Design and Intelligence Laboratory and Center for Biologically Inspired Design have studied interdisciplinary teams of designers and biologists engaged in semester long projects as part of a senior level Biologically Inspired Design course (BIOL/ME/PTEF/ISYE 4740) taught by CBID faculty.  The team has:

• developed consistent information processing models of biologically inspired design,
• compared the models with case studies of biologically inspired design practice as described in the literature,
• used the models to build computational tools for supporting education in biologically inspired design,
• applied the results to further refine the research. 

1. Co-evolution of Problems and Solutions

Problems and solutions in biologically inspired design typically co-evolve.  That is, while a designer’s initial understanding of a problem may result in the generation of an intermediate solution, the initial solution in turn may result in a better understanding of the problem.  In contrast to situations involving well-defined problem/solution spaces, bio-inspired design may be particularly useful in dealing with complex or ‘wicked’ problems. 

Salustri, Rogers and Eng in Designing as Balance-Seeking Instead of Problem-Solving argue that design problems often require approaches that a layperson typically does not associate with problem solving.  Interesting problems are often not static even after the requirements have been ‘frozen’.  Solutions may not completely resolve the problem, in that they may modify the problem condition. Algorithms or heuristics may not be available.  As problems become more complex, designers need to become more comfortable with “the idea of coevolution of problem and solution, where the act of solving a design problem illuminates the problem itself.”  Bio-inspired design may prove useful not just in providing specific solutions but also in supporting this design process.  Architects may have been particularly attracted to biologically inspired design due to the nature of the problem/solution space in which they work. 

2. Compound Analogies

Problem decomposition is an important aspect of design in general.  In biologically inspired design, analogical transfer of knowledge from biology to technology occurs at many levels of problem decomposition.  This leads to a process where many analogies are composed into a solution.  In contrast, many case studies of biologically inspired design that are reported in the literature are based largely on a single analogy, such as superhydrophobic paints based on the superhydrophobic properties of lotus leaves.  As more complex problems are tackled, the translation from biology into design may be equally complex, requiring broad insight into both the nature of the problem and the biological principles.

3. Problem-driven and Solution-based Design Processes

Biologically inspired design often engages two different processes with different starting points.  While problem-driven design starts with a technological problem and ends with a solution to it, solution-based design starts with a biological solution and ends with a technological problem and solution. 

Most people consider design to be problem-driven.  In contrast, the majority of case studies involving biologically inspired design are solution based, although this may be due to the current state of knowledge and expertise in the field.  Understanding both processes and how they complement each other may provide fruitful insights to both biologically inspired design and design in general.

4. Transfer of Functions and Causal Mechanisms

Analogical transfer from biology to technology in biologically inspired design typically pertains to knowledge of functions and causal mechanisms of biological systems.  Identifying appropriate biological systems is often based on knowledge of functions.  Units of knowledge that are transferred often pertain to causal mechanisms that result in accomplishment of these functions.  Thus, understanding and constructing models of biological systems in terms of their functions and causal mechanisms is a critical part of biological inspired design.

We are developing an interactive tool called DANE (for Design by Analogy to Nature) for supporting transfer of functions and causal mechanisms.  DANE provides access to a library of Structure-Behavior-Function (SBF) representations of biological systems.  An SBF model of a system informs the designer about the function of the system (or what the system does), the causal mechanisms in the system (or how the system works, called behaviors in SBF terms), and the structure of the system (or what the systems is made of).  A paper analyzing how students of the Fall 2009 CBID course used DANE will be published shortly (see Suggested Readings below).

5. Multimodal External Representations

Interdisciplinary teams of designers engaged in biologically inspired design typically use rich, intricate, multimodal external representations including drawings and equations.  These external representations play an important role in communication among biologists and engineers who often use different terminology and have different perspectives on design.  

Drawings and annotated diagrams in particular appear to help develop shared mental models of biological and technological systems that are accessible and meaningful to both biologists and designers.  These shared models can help communicate not only technical details but also outcomes and goals.  The diagram (based on a drawing made by a student project design team) links a wide range of biological information on the left with the final engineering solution on the right (for more information, see Findings from Cognitive Studies of Biologically Inspired Design in the November 2008 newsletter).

6. Situatedness in Information Worlds

Although design in general is characterized by relatively complete field-specific knowledge easily accessible to designers, biological information is typically found through textbooks, journals and expert sources that are difficult for designers to find, understand and apply.  Databases such as AskNature.org help bridge the gap, but it is not always easy to translate this material into useful design advice.  Researchers such as Dr. Li Shu of the Biomimetics for Innovation and Design Laboratory are developing natural language tools to search the existing biological knowledge base.

Our findings cover only a small portion of biologically inspired design, specifically the early conceptual design of engineering systems.  Although we expect that at least some of these findings are generalizable to other domains such as architecture and computing, this is yet to be empirically established.  Secondly, these results are preliminary: we will know more about their robustness and reliability only after we have used and confirmed them in many biologically inspired design situations.  Finally, we need to relate our work to the vast literature on design cognition in general.  Much work remains to be done!

Ashok Goel directs the Design & Intelligence Laboratory at the Georgia Institute of Technology.