Research Through Design

• Also called: RtD
• Category: Control › Human in the loop
• Tags: design-research, human-robot-interaction, prototyping, methodology, hci

A research approach in which designing and building prototypes is the main way of producing knowledge, with the artifacts and the reflection on making them treated as research results.

Definition

Research through design (RtD) is an approach to research in which the act of designing, prototyping and making is the central method for generating knowledge, rather than a step that follows it. The phrase comes from Christopher Frayling's 1993 paper Research in Art and Design, which distinguished research into design, research for design and research through design. In human-computer interaction it was developed further by Zimmerman, Forlizzi and Evenson, whose 2007 CHI paper described how designers can make research contributions by building artifacts that embody a proposed better future state.

In an RtD project, researchers frame a problem or opportunity, then iterate through concepts, sketches and working prototypes, documenting design decisions, alternatives considered and what was learned at each stage. The resulting artifacts are treated as carriers of knowledge: they make concrete an idea of how technology could be, and they can be deployed and studied with people. Contributions are usually generative rather than predictive, such as design qualities, frameworks, metaphors or annotated collections of artifacts that other designers can build on. Approaches such as annotated portfolios were proposed to make the knowledge in a set of artifacts explicit.

In robotics, RtD is most common in human-robot interaction, where researchers explore robot form, motion, materials and behavior through physical prototypes rather than starting from a fixed robot platform. It differs from engineering prototyping, where the prototype is a step toward a specified product, and from controlled experiments, where the aim is to test a hypothesis. RtD studies are judged by the rigor of their process documentation, the quality of reflection and the relevance of their contributions, rather than by statistical generalization. Critics note that this makes results harder to compare and replicate, which is why careful documentation of the design process is emphasized.

Why it matters

Some questions about how robots should look, move and fit into everyday spaces cannot be answered by testing existing robots, because the right designs do not exist yet. RtD gives researchers a recognized way to explore those questions by building, and it explains the structure and contribution of many design-led HRI papers that engineers might otherwise read as incomplete experiments.

Key applications

• Exploring robot form, materials and motion in human-robot interaction
• Designing social, ambient and domestic robots
• Speculative and critical design of future robot roles
• Developing design frameworks and metaphors for robot behavior
• Interaction design for wearable and assistive robotic devices

In practice

An HRI group wants to explore robots that share a living space without demanding attention. Instead of adapting an existing robot, they sketch concepts based on how plants move and grow, then build a series of prototypes using slow stepper-driven mechanisms, printed flexible leaves and fabric. Each prototype is placed in a home-like room, observed and discussed with participants, and the team records why each design choice was kept or dropped. The paper presents the prototypes, the design metaphors they arrived at and the open questions they raised, rather than a single performance number.

Common misconceptions

• Research through design is just building a prototype — the research contribution comes from documented, reflective iteration and the knowledge drawn from it, not from the artifact alone
• RtD studies are failed controlled experiments — they aim to generate ideas, frameworks and design knowledge, not to test a hypothesis statistically
• RtD is the same as engineering design or product development — engineering prototypes serve a specified goal, while RtD uses making to discover what the goal or the right questions should be

Who should know this

• Human-robot interaction researchers and designers
• Engineers collaborating with design researchers
• Reviewers of design-led robotics papers
• Students moving from engineering into HRI

Related technologies

• Rapid Prototyping
• Annotated Portfolios
• Participatory Design
• Speculative Design
• Field Deployment Studies

Related terms

• Wizard of Oz
• Godspeed Questionnaire
• Novelty Effect
• End-User Robot Programming

Related projects

• Webots: Cross-platform robot simulator with built-in models, sensors, and development environment
• Reachy Mini: Open-source desktop robot from Pollen and Hugging Face with an expressive 6-DoF head
• RAI: Vendor-agnostic agent framework for ROS 2 robots and physical AI
• Assistive Gym: Physics simulation framework for training assistive robots in caregiving tasks