A Computational Framework for Designing While Making
SM Thesis, MIT, Spring 2018
Terry Knight (MIT Architecture), Stefanie Mueller (MIT EECS), George Stiny (MIT Architecture), Maria Yang (MIT MechE)
In the wake of an increased accessibility of rapid prototyping tools in design practice, designers still face a series of challenges related to their use, one of them being the way in which they use these machines to actively explore, test and enhance design concepts. At the same time, the concepts of real-time interaction with computational fabrication tools and design exploration through physical prototyping are gaining impetus in computational design research and human-computer interaction.
Stimulated by these inquiries, the hypothesis of this thesis is that rapid prototyping tools can be used as tools for active design exploration and evaluation. I proposed Active Prototyping, a framework for embedding schemes for generative design into tools for interactive fabrication. To demonstrate the framework, I designed and built an apparatus integrating the operations of (a) physical prototyping, implemented through a hand-operated fused-deposition modeling (FDM) prototyping system, (b) recording of designer actions, implemented through a system for position sensing and storing actions into shapes, and (c) visual exploration of possible solutions, developed through an interface for implementing spatial transformations on the stored forms.
Active Prototyping contributes to ongoing inquiries in the areas of computational design, design research and human-computer interaction that investigate how the human ability for creativity and thinking while making can be augmented through the computational capabilities of information storage and processing.
Active prototyping is introduced at a time and in a technological reality where the way that designers learn and improve their artifacts through in-situ, real-time prototyping practices (tool- using) is being widely investigated. At the same time, it is inspired by new systems for integrating prototyping machines into the design process (tool-making) that are mostly developed within the fields of engineering and human-computer interaction. A comparison between current tool-using and tool-making approaches indicates that analysis of what makes a prototyping procedure creative and invention of new techniques for efficient control of prototyping tools are deployed as two remote enterprises. Based on an analysis of current divergences and convergences between these two directions, I propose Active Prototyping as a framework for integrating tool-making agendas with tool-using practices. This observation calls upon the integration of systematic methods for design exploration into the workflow of prototyping tools. In my proposal, I introduce these methods through generative design frameworks. While research on prototyping tools and activity demonstrates how design can be enhanced through physical making, frameworks for generative design propose a computational, representation-based way for enhancing design. The question that then rises is: how can prototyping tools be used in a generative way? Active Prototyping takes advantage of the main conceptual and technical tools of generative design methods such as rule applications, solution enumeration and selection in order to answer this question.
The proposal of developing a human-machine fabrication system based on research coming from these diverse backgrounds does not reflect an intention for creating a hybrid assembly between them, or one for importing concepts from one to the other. In the case of prototyping research, the premise is that design research and human-computer interaction look into tools from opposite sides – design research focuses on tool-using whereas interactive fabrication focuses on tool-making–, and that, within a design prototyping context, the one can be used to support the function of other. The method of this thesis is centered on the integration of frameworks that support both operations of action analysis and translation into computer input and output. This is where generative methods come to contribute. I use generative methods as the conceptual tool for translating results gathered through prototyping activity into computational tool inputs. On the basis of the Active Prototyping framework is the identification of actions that can be recorded and extracted into information that can be visually represented. To explore this hypothesis, Active Prototyping deploys the following operations:
Physical control of prototyping tools that are connected to a computer system,
Action recording through continuous calculation of the position of the tool while it is being operated by the user,
Exploration of different design alternatives via the application of generative design rules to the recorded inputs,
Machine feedback in the form of automation and guidance of manual designer actions.
I propose Active Prototyping as a system for aiding designers to track their actions as they work on physical prototypes, and therefore to explore, archive and evaluate the design possibilities that these actions relate to. The technical implementation of this framework demonstrates how these operations relate to each other in action and within a physical context. These aspects have both a physical and a digital expression - physical in the sense of analyzing actions, and digital in the sense of recording, retrieving, processing them and bringing them back into the physical prototyping process. This duality creates the need for two parallel operations: a manual operation that allows designers to perform physical prototyping actions and a digital operation of tracking actions, extracting data from them, storing and processing that data, and making them available to designers in real time.
I develop Fabcorder as a prototyping device that implements the operations introduced by the Active prototyping Framework. The device consists of a work table attached to a hand-operated tool for material extrusion. A series of systems are combined to integrate the Active Prototyping operations:
Physical control of prototyping tools is implemented through a two-state prototyping system for material extruding and gesturing.
Action recording is implemented through a position tracking mechanism that is attached to the prototyping tool. The mechanism consists of three rotary sensors that are connected to each other through a system of taught strings. The sensors convert the angular position and motion of the tool to a digital signal. The digital signal produced from each sensor is communicated to an interface for shape calculation, display and storage. While the position tracking mechanism records movement in three dimensions, the work table can move across the vertical axis. The motor system that activates the table communicates with the system that tracks and records tool position and movement, and therefore designer actions can be detected in three-dimensional space. By communicating the tool state described in (i) to the shape display interface, the system can record and represent not only the position of the movement in space, but also its effects to the physical object that is being prototyped.
Design Exploration is implemented through a visual system for manipulating recordings and letting the user save, retrieve, re-use and re-design them as rules for generative design. Examples of such projections are transformations of an extruded profile which, stacked on top each other, form a three-dimensional prototype. The purpose of these manipulations and their projection is to define a range of design possibilities and outcomes after every step of the prototyping process, and to aid the designer choose among these possibilities in her next physical action.
Machine feedback on the next steps of the prototyping process is proposed through two methods that are integrated into the action recording and the visual display systems.
In this thesis, I proposed Active Prototyping as a conceptual framework and as a technical workflow for integrating a generative approach into rapid prototyping practices. I proposed that, in order to serve as a better medium for expression, prototyping must be active besides rapid: active in the sense of integrating designer actions and decisions that are being developed during the prototyping process. On the center of this framework and its associated apparatuses was recording, which consists the main contribution of this thesis, and shapes its agenda for future work:
recording as a result, referring to the a posteriori collected data of a design process and how they can be analyzed and reused,
recording as a process, referring to the data collected and processed while a design and prototyping process is being developed, and how they are used to enhance the process itself,
recording as a research tool, referring to how, not only as designers but also as tool-makers and design researchers, we can learn from the way that other designers and makers design, prototype, explore and materialize their ideas.
Through Active Prototyping, recording prevailed as the main mediator between human actions and machine feedback. This thesis took a first step in exploring this relationship, its potentials and its dependencies on the technologies that are being used. The conclusion from this discussion is that, in a human-machine collaboration framework, what can be recorded can be translated into computer input, and consequently be computationally augmented. The challenge, in future work, lies in how we perceive and design what is to be recorded, and how this perception can enrich what we want to achieve through this recording.
Through Active Prototyping, I introduced a novel framework for human-machine collaboration that takes advantage of human dexterity and the ability for thinking through making and the computational capabilities of information processing and storage. Moreover, I took a first step in shaping future technical agendas for more human-centered and creative uses of tools for prototyping and fabrication.