BIKAS: Bio-Inspired Knowledge Acquisition and Simulacrum—A Knowledge Database to Support Multifunctional Design Concept Generation

Bioinspiration is a constant source of inspiration for many engineers, scientists, and designers to generate innovative and unique design solutions for complex engineering problems. Bio-inspired design (BID) is an approach that uses analogies to biological systems to extract innovative solutions to solve difficult or complex engineering problems [1,2]. Ever-increasing human demands and environmental problems require innovative solutions. For example, sustainable products are required to reduce the overall environmental impact, and innovative solutions are required to tackle emerging complex engineering problems such as water treatment, waste management, etc. Bioinspiration is a source to provide innovative solutions to such problems. However, the traditional product development processes depend on the incremental improvements of single products [3]. Therefore, seeking a radical, innovative, and agile product development process is imperative to satisfy current human demands and needs. Bio-inspiration or Bio-inspired design is one of the emerging creativity and innovation methods [4] that aids in developing innovative and efficient products. Biological systems are multifunctional as they survive by simultaneously performing various tasks [5]. Moreover, biological systems perform various functions by reducing the number of parts, thus reducing material and energy consumption [6-8]. Biological systems offer highly efficient solutions for various emerging complex engineering problems, for example, effective energy absorption [9], effective antibiofouling and self-cleaning [10], heat isolation, heat exchangers, vibration control, etc [11]. Nevertheless, repositories of biological features exhibiting various functions are required to develop methods and frameworks aimed at utilizing nature principles to achieve innovative conceptual designs for multifunctional bio-inspired designs. The following sub-section provides the details of the existing repositories of biological information.

Biological information, representation, and analysis are essential for understanding the biological system or a feature for emulating its functionality. Repositories were developed to enhance the understanding of biological systems, biological features, and the functions exhibited by them. For example, Ideas Inspire is a software-based repository representing biological systems using the SAPPhIRE model of causality [12]. DANE is another repository of biological systems represented using function-based approaches such as Structure-Behaviour-Function (S-B-F) [12]. BioTRIZ is a unique approach developed to integrate biological knowledge and TRIZ problem-solving technique. The biological information is represented using a descriptive approach and interactively positioning the biological strategy in the BioTRIZ contradictory matrix [13]. AskNature is an interactive software-supported repository of biological information represented by using a functional, hierarchical taxonomy called Biomimicry Taxonomy [13, 14].

Table 1 presents the existing repositories of biological information, the level of abstraction, and the approach used for the representation of the biological system. Ideas Inspire and DANE abstract the biological analogy as a whole system. Both repositories use functional modelling technique. Functional modelling is a system-level representation in the form of a block or tree diagram that describes the process flow between components, for example, matter, energy, or signal. This approach is often associated with the functional decomposition of the overall function into sub-functions [15]. On the other hand, AskNature and BioTRIZ have their own unique modes of representation of biological analogies. AskNature uses Biomimicry Taxonomy and BioTRIZ uses descriptive representation integrated with parameters such as substance, structure, time, space, energy, field, and information adaptation [16]. However, to reduce the complexity in abstraction of a functionality in a biological organism and enhance the rapidity in ideation, a unique biological knowledge database called BIKAS is introduced.

This paper presents the knowledge database BIKAS: Bio-inspired Knowledge Acquisition and Simulacrum of biological systems exhibiting various functions represented according to their biological features, characteristics, and function as a combination of integrated structure and structural strategy. BIKAS is developed by analysing over 50 case-based bio-inspired examples from the literature.

This paper is organized as follows: Section 2 provides a description of the biological features, their characteristics, the integrated structure, the structural strategy, and the BIKAS. Furthermore, it presents data visualisation of the knowledge database. Section 3 presents the construction of BIKAS and criteria for evaluating databases. Section 4 elaborates the utilization of the knowledge database in the xDID model. Section 5 presents conclusions.

BIKAS is built based on the case-based bio-inspired design examples in the literature. The biological systems are represented according to the biological feature, its characteristics, and the function exhibited by the feature as a combination of the integrated structure and the structural strategy. Table 2 presents the abstraction level, mode of representation, and methods and model associated with the BIKAS. The biological features that exhibit various functions are mapped to their structure, and structural strategy. The level of abstraction of a biological entity in the case of BIKAS is at the biological feature level or in other words at the component level.

The following points describe each component of the BIKAS.

• The Meta-Level Embodiment function represents the function exhibited by the biological feature at the embodiment function level (at the physical structure level) [15, 17].

• The Biological features refer to the morphological and anatomical features observed in plant and animal kingdoms.

• The Capitalize feature characteristic represents the feature's appearance, apparent form, or physical trait. For example, body/skin texture, hard outer composite covers like outer composite plates, outer composite tiles, outer composite shells, cellular structures, etc.

• The Capitalize structure represents the physical description of a multiscale structure (e.g., Micro/Nanostructure, Macrostructure, and the presence of wax layers on the Structure, etc.) [15].

• The structural strategy represents the integrated structural configuration (e.g., arrangement of the micro/nanostructure, packing of the micro/nanostructure, orientation of micro/nanostructure, symmetry, asymmetry, or patterns of tessellations, etc.) and changes in the structural configuration due to stimulus. Stimulus are by the other interacting elements connected to the structure (e.g., erection of scales, change in skin compliance, etc.) [15].

Figure 1 shows the schematic representation of the biological feature its characteristics and the function exhibited as a combination of its integrated structure and structural strategy.

The repel water function exhibited by the Butterfly wings shown in Table 3 is an example of BIKAS. The meta-level embodiment function to repel water exhibited by the butterfly wings is a combination of the integrated structure: a directional hierarchical arrangement of the nano-tips on the micro projections [18] and the structural strategy: orientation of the structure (on the wing) in the downward direction [18].

Figure 2 shows the schematic sketch of the nano-tips and hierarchical microstructure of the butterfly wings. The following sub-section provides the BIKAS. In BIKAS, the key structural strategies are represented in bolditalics and where are braces?

BIKAS is utilized by Domain Integrated Design (DID), a newly developed bio-inspired design method for multifunctional conceptual designs. DID differs from other multifunctional bio-inspired design methods by following a unique approach of classification, mapping, and representation of biological features. The biological features responsible for exhibiting various functions are classified by their feature characteristics to their respective geometric designations called Domains. The Domains are Surfaces, Cellular Structures, Cross sections, and Shapes. Furthermore, the classified biological features are mapped to the tissues they originate from. In addition, the function exhibited by the biological features is represented as a combination of the integrated structure (often multiscale) of the biological feature and its structural strategy. A detailed comparison of the DID method and existing multifunctional bio-inspired design methods is reported in [16]. Expandable Domain Integrated Design (xDID) model is an extension of the proposed DID method to facilitate the classification of complex biological features that exhibit various functions that cannot be described according to the current definitions of the DID method. Furthermore, the xDID model facilitates the creation of micro-Domains within the same Domain for accurate description and classification of the biological feature. The classification, mapping, and representation of the xDID model with case studies are detailed in [15]. The next section presents the knowledge database that is BIKAS. The biological analogies and their information presented in BIKAS are manually curated. Currently, around 50 biological features are analysed and represented in BIKAS. The knowledge database is expected to expand as more biological entities are analysed and represented according to the mapping process in BIKAS. Furthermore, expansion of the BIKAS can be used for implementing natural language processing algorithms for an automated population of the database.

Table 4 describes the meta-level embodiment functions exhibited by biological features, their characteristics, integrated structure and the structural strategy.

This section provides the visualization of BIKAS that comprises the biological features performing various functions. Figure 3 presents the biological features, their characteristics and the function exhibited by the features. This visualization provides an insight that the biological features of distant biological organisms performing a similar function.

In literature, this phenomenon is described as convergent evolution, where distant biological systems exhibit the same function in radically different ways. For example, the stem cross section of the bamboo tree exhibits the function of absorb energy [9]. Likewise, the elytra of a beetle forewing exhibit similar function of absorb energy [35]. Another observation from Figure 3, is that some feature exhibit more than one function for example, the reduce aerodynamic and fluidic drag function is exhibited by shark [46] and dolphin skin [47] and the resist bending and absorb energy function of the Toucan and Hornbill beaks [45].

The Figure 4 shows the measure of biological feature characteristics that were mimicked to achieve a particular function in various case-studies. As shown, the biological feature characteristics Body/Skin texture (Animals) was mimicked the most followed by the Body/Skin texture (Plants). As BIKAS expands, more features will be added and expected to provide a more comprehensive and more detailed and user interactive approach in utilizing the knowledge database.

BIKAS is a unique knowledge data that involved a manual search, acquisition, analysis, and curation. For structured knowledge bases and repositories that involve creation of representation modes for understanding biological phenomenon require human interaction, initiation and curation [13]. A construction flowchart is necessary to describe the creation of BIKAS. Construction flowcharts enable to reproduce, populate, and enhance the understanding of biological feature of a biological entity. Figure 5 shows the schematic of the construction flowchart. The construction flowchart of BIKAS comprises of three phases namely, Acquisition phase, Analysis Phase, and Interpretation phase. The following are a detailed description of each phase.

• Acquisition phase: This phase comprises of acquiring the bio-inspired case-studies and research articles that describe biological features, their functions, principles, replication, manufacturing, and application of biological structures.

• Analysis phase: This phase comprises segregation of biological features, feature characteristics, structure, and structural strategy of the biological feature.

• Interpretation phase: This phase comprises of interpretation and structuring the acquired and analysed data. Mapping of function-feature-structure-structural strategy takes place in this phase.

All the three phases involved in a manual search, acquisition, analysis, and curation. The acquired data is stored as open-source database using an interactive web-based data visualisation tool. The data sources for data acquisition are mainly research articles and case-studies of biologically inspired structures searched in popular sources such as Scopus, Google scholar, and special issues and journals in bio-inspired design. Often biological-inspired case studies may not have all the necessary information for the construction of BIKAS. For example, a detailed description of the structure and the structural strategies (configuration of the structure). In such cases, more articles regarding the biological system are researched and analysed for the construction.

The effectiveness of qualitative databases that involves in analogical transfer such as biological data for engineering design depends upon how well the biological data is understood [70]. For bio-inspired design knowledge databases that only has mode of representation of biological features, the evaluation is based on how well the biological information is abstracted and understood, and how well the information is used to build a solution. Two such modes of biological representation have been evaluated so far. Those are namely, SAPPhIRE [70] and S-B-F (Structure-Behaviour-Function) [71]. The evaluation is a comparative study on SAPPhIRE representation of biological information and a generic text-image representation through a series of questionnaires provided to the designers [70]. Similarly, a comparative study between S-B-F representation, text-only and text-graphic-tabular representation was carried out through a series of questionnaires [71]. Table 5 presents the evaluation criteria for evaluating repositories for bio-inspired design. As shown, these modes of representations do not have a method associated with them. Such repositories without an associated methods are evaluated through a comparative study.

However, BIKAS is a unique database of biological information represented as a mapping between function-feature-structure-structural strategy. In addition, a method classifying the biological features according to their characteristics into their geometric designations called Domains is associated with the database. The principle is combination of features from different Domains results in multifunctional conceptual designs. BIKAS and DID are unique mapping and classification of biological features. The evaluation criteria for such a representation and its associated classification method are the design and simulation-based evaluation of early-stage embodiment designs.

The effectiveness of such a complex mapping and classification of biological features is verified by designing unique bio-inspired multifunctional early-stage embodiment product concepts and validating them using computer simulations. Several case-studies were published that evaluates the effectiveness of the knowledge database and the method. For example, multifunctional bio-inspired painless sutures [53], multifunctional bio-inspired effective heat transfer and low pressure drop structures for aerospace applications [72], multifunctional bio-inspired non-pneumatic tire design for effective friction management and impact resistance [73].

xDID model works on classifying the biological features into their geometric designations, mapping the classified biological features to their respective tissues from which they originate, and representing the function performed by the biological features as a combination of the integrated structure and its structural strategy. In this model, firstly, the biological features are classified by their feature characteristics to their respective geometric designations called Domains. The Domains are namely Surfaces, Cellular structures, Cross-sections, and Shapes. Domains represent different biological features performing various functions, mapped to their tissues with a common geometric designation. Figure 6 shows the representation of the repel water function exhibited by the biological feature, which is the micro/nano projections of rose petals as a combination of its integrated structure and structural strategy. The integrated Structure is the rounded micro-bumps, and its structural strategy is the microstructure's random arrangement and dense packing. In addition, the biological feature is mapped to the biological tissue from which it originates.

The descriptions of the classified Domains was extensively discussed in [15]. However, this sub-section provides a brief description of the Domains. The Surfaces Domain comprises the biological features that can be described with feature characteristics such as animal and plant body/skin textures, structures that form the body coat, for example, wool, hair, etc., and the elements that include the hard outer and sandwiched cover such as shells, plates, and tiles. The Cellular structures Domain comprises the biological features described as porous prismatic and foam structures arranged in various tessellated patterns, namely periodic, stochastic, and hierarchical, with feature characteristics such as periodic tessellations with unary, binary, ternary or quaternary connections, stochastic tessellations with Poisson distribution, Voronoi or crystal growth patterns, and finally hierarchical tessellation with branching, nested or overlaid connections [11]. The Cross-sections Domain comprises the biological features with the characteristic of a cross-section of the body or a part of the body. Cross-sections are described as shapes obtained by cutting a body along the lateral and longitudinal axis. The shape Domain comprises the biological features with characteristics of the body's external form or contour, such as a boxfish shape [56] for drag reduction or a part of the body's external form or contour that exhibits a specific functionality.

Figure 7 shows the classification of biological features to their respective geometric designations (Domains) based on their biological feature characteristics for the xDID model. The shapes and crosssections are considered sub-Domains of surfaces and cellular structures because the function exhibited by biological features in these Domains are due to the combination of different biological characteristics or different biological tissues. The cross-sections and shapes sub-Domain have an associated biological feature characteristic belonging to Surfaces Domain or Cellular structures Domain. Figure 8 shows the comprehensive visual representing the biological feature, the function the biological feature exhibits, and the Domain to which the feature is assigned.

However, biological features are highly complex structures. They cannot be described by just one characteristic, making it highly impossible to systematically classify biological features into a finite number of Domains. xDID facilities addition of new Domains for the features that cannot be described according to the definitions of the current study. In addition, xDID offers the creation of micro-Domains for more precise description and classification of the biological feature by its feature characteristics. As shown in Figure 9 [15], the Surfaces Domain can be sub-divided into micro-Domain that comprises of biological features with feature characteristic of hard outer cover such as sandwiched plates, tiles, and shells. Similarly, the cellular structures Domain can be subdivided into micro-Domains by the type of connections such as beam based connections or face based connections [15].

In xDID, the functionality is seen as a combination of the integrated structure and the structural strategy. The proposed Meta-level design factors are the Domain-specific structural strategies observed from the literature case studies that influence a biological system's functions. The abstracted Meta-level design factors aid as a filter to extract the features and their functions from BIKAS that are influenced by a similar biological structural strategy.

Table 7 presents the list of Meta-level design factors for various Domains and sub-Domains proposed in xDID based on the analysis of bio-inspired design cases. The Meta-level design factors are filters to extract features and functions within the same Domain that share a similar structural strategy. This extraction aid designers providing with different biological features and their functions that have a similar strategy thus enhancing the rapid selection and emulation of the biological features for rapid ideation process. Figure 10 shows the application of Meta-level design factor as a filter on the knowledge database to extract biological features that exhibits a particular function with a similar structural strategy.

Table 8 shows an example of the application of the Meta-level design factors, where two functions one is repel water function by lotus leaf and the other is resist erosion function by scorpion skin, both having the similar structural strategy that is random arrangement of the micro/nano projections.

To develop multifunctional bio-inspired conceptual designs and methods a detailed acquisition, analysis and representation of the biological features are required. BIKAS is a knowledge database developed based on the case-based bio-inspired design examples from the literature. BIKAS aids in developing multifunctional bio-inspired design concepts and methods. xDID model is one such method developed utilizing the knowledge base. The unique classification, mapping, and representation of biological features will bridge the gap between biology and engineering fields and enhance biologists’ collaboration that results in agile innovation. Unlike the existing multifunctional design methods, the xDID method enhances the rapid ideation process. However, the current approach has certain limitations. Initial DID method emphasised on assigning the biological features to four Domains namely Surfaces, Cellular Structures, Cross-sections, and Shapes. But in the vast Plant and Animal kingdoms complex functional biological features exists that cannot be assigned to a Domain based on the current definitions. To tackle such a complex situation xDID approach is introduced that facilitates the creation of micro-Domains within a Domain and new Domains.

BIKAS is a unique mapping of function-feature-structure-structural strategy in association with DID method that classifies features into their geometric designations called Domains results in generation of multifunctional bio-inspired conceptual and early-stage embodiment designs. Moreover, the abstraction of biological entities is at its feature level or component level. This enhances the rapidity in ideation by replicating similar structures. However, BIKAS lacks in a systemic-level analysis of a biological entity. Nevertheless, BIKAS can be integrated with other system-level approaches for representation to enhance the understanding of the biological entity both at system and component levels.

This paper presented a knowledge database of different biological features performing various functions. Furthermore, functions exhibited by the biological features are represented as a combination of integrated structure and the structural strategy. Additionally, xDID model for generation of multifunctional conceptual design is discussed. Finally, Meta-level design factors are abstracted, which are the Domain-specific factors that act as a filter to extract the biological features and their functions within the same Domain that have a similar structural strategy. Meta-level design factors aid designers in filtering and sorting the biological features that have a similar structural strategy for rapid selection and emulation. Effective utilization the BIKAS and xDID model is achieved by populating the existing knowledge database and addition of new Domains.

The next step is to develop an open-source tool for designers, biologists, and scientists to utilize the xDID model.

Pavan Tejaswi Velivela conducted the research, categorization, and analysis. Yaoyao Fiona Zhao supervised the study and edited and revised the manuscript.

This research work is supported by the Natural Sciences and Engineering Research Council of Canada Discovery Grant RGPIN-2018-05971 and MEDA (McGill Engineering Doctoral Award).

The authors declare no conflict of interest.