Tremor Reduction for Accessible Ray Based Interaction in VR Applications

TL;DR

This paper proposes a low pass filter to reduce tremors in ray-based VR interactions, enhancing accessibility and accuracy for users with fine motor challenges, supported by a user study and design analysis.

Contribution

It introduces a novel filtering algorithm for tremor reduction in VR ray interactions, improving accessibility and user experience.

Findings

The filter improves interaction accuracy and reduces user frustration.

Users with tremors experienced more precise control with the filter.

The approach supports accessible VR design by accommodating fine motor limitations.

Abstract

Comparative to conventional 2D interaction methods, virtual reality (VR) demonstrates an opportunity for unique interface and interaction design decisions. Currently, this poses a challenge when developing an accessible VR experience as existing interaction techniques may not be usable by all users. It was discovered that many traditional 2D interface interaction methods have been directly converted to work in a VR space with little alteration to the input mechanism, such as the use of a laser pointer designed to that of a traditional cursor. It is recognized that distanceindependent millimetres can support designers in developing interfaces that scale in virtual worlds. Relevantly, Fitts law states that as distance increases, user movements are increasingly slower and performed less accurately. In this paper we propose the use of a low pass filter, to normalize user input noise, alleviating fine motor requirements during ray-based interaction. A development study was conducted to understand the feasibility of implementing such a filter and explore its effects on end users experience. It demonstrates how an algorithm can provide an opportunity for a more accurate and consequently less frustrating experience by filtering and reducing involuntary hand tremors. Further discussion on existing VR design philosophies is also conducted, analysing evidence that supports multisensory feedback and psychological models. The completed study can be downloaded from GitHub.