illumotion: An Optical-illusion-based VR Locomotion Technique for Long-Distance 3D Movement

Abstract

The Locomotion Trilemma

VR locomotion has a fundamental trilemma: steering (continuous movement via joystick/controller) preserves presence but induces cybersickness through vestibular-visual conflict; teleportation (instant jump to target) eliminates cybersickness but breaks presence and spatial orientation; room-scale walking is ideal but constrained by physical space. Every new locomotion technique is an attempt to triangulate a better position within this trade space. illumotion's bet is that the trilemma itself is partly an artifact of framing — that "moving through a space" and "manipulating what you see" can be perceptually equivalent if the optical transformation is designed correctly.

How illumotion Works

The technique is conceptually borrowed from touchscreen manipulation. On a phone, you pinch to zoom into a photo — your fingers don't move through the image, but the visual result is identical to moving closer. illumotion applies this logic to 3D VR scenes through three maneuvers:

• Zooming: the user pinches both hands and pulls them apart or together, producing forward/backward movement through the scene. The optical analogy is changing the field of view while keeping the scene in frame.
• Dragging: a single-hand drag translates the view laterally, equivalent to panning across a 2D image but solved in 3D screen space.
• Pulling/pushing: a two-hand vertical gesture that shifts the viewpoint up or down.

The critical technical challenge is solving the motion parameters in screen space. Unlike 2D photo manipulation where the image is flat, a 3D VR scene has depth — zooming toward a nearby object should produce slower movement than zooming toward a distant mountain, or the user feels either out of control (near objects rush past) or stuck (distant objects barely move). illumotion addresses this with a depth-aware speed model: the optical flow at the center of the user's gaze determines the movement rate. Look at something far away, and you move faster; look at something close, and you decelerate. This gives the user implicit speed control through gaze targeting, without requiring explicit throttle input.

The Illusion That Matters

The technique's name is well-chosen. The user is not actually moving — the viewpoint is being translated by the system — but the optical transformation mimics the retinal flow of real movement closely enough that the brain accepts it. This is the "optical illusion" at work: by anchoring the interaction to hand gestures that feel like manipulating a scene (rather than steering a body), illumotion may sidestep the vestibular-visual conflict that causes cybersickness. The brain has no evolutionary expectation for what "pinching and pulling a world" should feel like in the inner ear, so it doesn't flag the mismatch as poisoning.

If this hypothesis is correct — and the cybersickness data from the user study supports it directionally — then the design principle generalizes beyond this specific technique: locomotion methods that frame movement as object manipulation rather than self-motion may inherently produce less sickness, regardless of the specific gesture set.

The User Study

Sixty-six participants compared illumotion against both steering and teleportation on multiple dimensions: performance (task completion time, accuracy of target approach), presence (standard presence questionnaires), usability (SUS), user experience (UEQ), and cybersickness (SSQ). The scale is notable — 66 participants is large for a VR locomotion study, where 20–30 is more typical.

The results are unusually clean: illumotion outperformed or matched both baselines on every measure. It was faster than steering for target-reaching tasks. It produced higher presence scores than teleportation (as expected — continuous movement preserves spatial context). It scored higher on usability than both. And crucially, it produced lower cybersickness scores than steering, while matching teleportation on this dimension. In other words, it captured the benefits of both established techniques — steering's presence, teleportation's comfort — without paying either's primary cost.

What to Be Skeptical About

The 66 participants were presumably VR-naive or moderately experienced (the paper reports demographics but the generalization to expert VR users, who may have different sickness thresholds and locomotion preferences, is uncertain). The tasks were target-reaching in open scenes — performance in cluttered environments, narrow passages, or tasks requiring simultaneous object manipulation is untested. And the long-term retention question is open: do users who adopt illumotion as their primary locomotion method eventually develop cybersickness after extended sessions, or does the optical-illusion framing provide durable protection? The study tested a single session; longitudinal data doesn't exist.

The depth-aware speed model is both a feature and a constraint. Gaze-targeted speed control works well when the user is looking where they want to go, but breaks down in scenarios where the user needs to look elsewhere while moving — checking a side path while advancing, monitoring a companion, scanning for threats in a game. In these cases, the speed becomes unpredictable, which may itself induce discomfort.

Finally, the technique requires both hand tracking (for the pinch/drag gestures) and a depth buffer (for the speed model). On current hardware, this means it works best on headsets with inside-out hand tracking (Quest series, Apple Vision Pro) and may not translate cleanly to controller-only systems. That's an adoption constraint, not a research flaw — but it bounds the technique's near-term reach.