A look at the iterative design of a touchless interface
A touchless interface lets people control a screen or installation without touching a surface. Gestures, proximity and motion sensing can make an exhibit feel immediate, but the visible gesture is only one part of the interactive system.
Iterative design turns uncertain ideas into testable behaviours: sketch, prototype, observe and revise. For an online showcase such as ITP 30 Show, that process gives experimental art and media projects a clear account of how their interactions developed.
From touch to spatial intent
A swipe on glass has a clear beginning and end. A swipe through air is ambiguous: is the user selecting, scrolling or signalling to another person? Touchless systems therefore depend on context as much as recognisable gestures if the user experience is to feel natural.
Before choosing sensors, define what each gesture means, when it starts, how it ends and how it is cancelled. A raised hand might pause an illustrated catalogue, while two hands moving apart could adjust scale. These rules form an interaction contract that the hardware and interface must share.
Why prototypes expose hidden assumptions
A prototype should answer a question rather than merely resemble the finished work. Paper cut-outs, recorded gestures and a camera feed mapped to one action can reveal whether people understand a movement before complex rendering is added.
Track the delay between movement and response, along with missed activations and accidental triggers. If a visitor waves repeatedly, the system has not communicated its state. Each iteration should test one important assumption, such as threshold placement, and preserve the previous version for comparison.
Matching sensors to local conditions
Optical tracking can estimate joints; infrared proximity sensors detect presence; ultrasonic sensors estimate distance; radar can work in darkness but may provide less precise body data. Each approach carries assumptions about range, movement and privacy, so sensor choices should follow the experience rather than lead it.
Local conditions matter. Strong daylight in Brisbane or Perth can reduce contrast for some optical sensors, while busy railway platforms in Sydney introduce crowds, reflections and background movement. A sensor that works in a quiet studio may therefore fail in the everyday settings where the interface is meant to operate.
Designing for Australian everyday use
Cashless payments are routine in Australian shops, cafés and grocery stores. Many people already tap a card or phone for a coffee, while Opal in Sydney and Myki in Melbourne reward quick taps for public transport. That familiarity is useful, but contactless payment and touchless control solve different problems; users do not automatically expect mid-air gestures at a ticket machine.
Camera-based prototypes also raise privacy questions. Where the Privacy Act 1988 and Australian Privacy Principles apply, collection and use must be justified, transparent and limited to stated purposes. Gesture detection should not be confused with consent to record. Make cameras visible, process movement on the device where possible and avoid storing identifiable footage by default.
Accessibility requires equal access, not novelty at any cost. Under disability discrimination law, including the Disability Discrimination Act 1992, public-facing systems may need reasonable adjustments. Some users cannot sustain a raised-arm gesture, so provide buttons, voice control or dwell activation alongside the touchless mode.
Liveness, feedback and responsive systems
A system feels lively when its response belongs to the present interaction. The ITP 30 Show journal offers a reflection on liveness, which connects this quality to the experience of participating in an interactive work rather than merely viewing a fixed presentation.
Liveness differs from constant motion or realistic three-dimensional graphics. It emerges when a person can recognise cause and effect: a hand approaches, the system acknowledges it, and the next action follows predictably. Random reactions may look busy while remaining difficult to understand.
Feedback should therefore be immediate, legible and proportional. A soft outline, low-volume sound or subtle change in scale can confirm that a gesture registered without demanding another movement. Low latency matters too, since delay separates the person’s intention from the system’s response.
Recommendations for documentation and iterative testing
An interactive project needs documentation that preserves behaviour, not just appearance. Practical lessons in web documentation explain how descriptions, demonstrations and credits can give an online audience a reliable sense of a project that cannot be handled in person.
Record the setup, input method, operating conditions and known limitations alongside a short demonstration. Keep the gesture thresholds, fallback controls and privacy choices visible. A change log makes the iterative design legible by showing which assumptions produced each revision.
- Define every gesture’s trigger, release and cancellation rules before coding.
- Test paper or video prototypes to see whether the movement communicates its intended action.
- Measure response delay, missed activations and unintended triggers in realistic conditions.
- Repeat trials in bright daylight, low light and crowded backgrounds.
- Provide accessible alternatives that do not require mid-air movement.
- Record consent, minimal data, setup details and changes between versions.
For the web, pair written instructions with captions, still frames and a demonstration showing the interaction unfolding. Mark which parts depend on hardware, browser or location, so future visitors can distinguish a deliberate design choice from a temporary constraint.
Build a small test and share what you learn
A useful final round brings the prototype back to its setting. Invite participants to approach naturally, ask them to think aloud, and watch where they hesitate. Consent should cover observation, recording and later publication, and a familiar control should remain available when the gesture system is unavailable.
Choose one meaningful gesture, build the simplest version that can detect it, and run one recorded test with a collaborator this week. Share the result through a project description, process notes and a clear demonstration of what changed.