ExtruFace

Overview

ExtruFace was an applied research project developing the concept and prototype of a gesture-controlled interface for industrial extrusion machines. I was involved across all phases: requirements analysis, design, implementation, and final acceptance.

The Problem

Extrusion machines are large, heavy industrial equipment used for pipe forming. Operators control them from varying distances and routinely wear protective gloves, which makes the existing touchscreen solution impractical for much of a normal shift. The question the project set out to answer was whether a gesture-controlled interface could replace or supplement it, and what that would need to look like to be actually usable on a factory floor.

From Interviews to Interaction Model

The design process started with operator interviews, not sketches. Understanding how operators actually move around the machine during a production run, which controls they reach for frequently and from where, and which situations require quick access versus careful deliberation: all of this shaped the core design decision the interface needed to behave differently depending on how far the operator was standing from the screen.

This led to a two-mode model. In far mode, triggered when the operator is several meters away, only the controls most commonly needed at distance are shown, and input happens through body gestures (arm poses) detected by a Microsoft Kinect. The far-mode gesture set was developed together with operators, drawing on research from my bachelor's thesis on gesture design.

In near mode, when the operator approaches the screen, the interface expands to its full set of controls and switches to fine-grained touchless hand gestures via a Leap Motion Controller. The transition between modes happens automatically based on distance tracking.

Menus in both modes use circular layouts, minimizing the travel distance between options: a direct application of Fitts' Law to gesture input, where moving the hand across a large distance is expensive in a way that moving a mouse cursor is not.

Prototype and Evaluation

I built a functional web prototype demonstrating both interaction modes, which was tested with real machine operators in a laboratory setting. A touchscreen fallback was included as a safety net for situations where gesture control was impractical.

The evaluation confirmed that the adaptive concept addressed the core problem effectively. Operators could control the machine from distance without removing gloves, and the near-mode hand gesture interaction held up for detailed tasks. The prototype was handed over as a source of ideas for the client's further development of their existing system.

Outcome

The project demonstrated that touchless gesture control is viable in an industrial machine context, provided the interface adapts to the operator's physical position rather than expecting a fixed interaction posture. The central contribution was the transition mechanism between near and far mode: automatic, distance-driven, with different gesture vocabularies for each. It held up in testing with real operators.