Predicting Mid-Air Interaction Movements and Fatigue Using Deep Reinforcement Learning

Noshaba Cheema, Laura A. Frey-Law, Kourosh Naderi, Jaakko Lehtinen, Philipp Slusallek, Perttu Hämäläinen

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

54 Downloads (Pure)

Abstract

A common problem of mid-air interaction is excessive arm fatigue, known as the "Gorilla arm" effect. To predict and prevent such problems at a low cost, we investigate user testing of mid-air interaction without real users, utilizing biomechanically simulated AI agents trained using deep Reinforcement Learning (RL). We implement this in a pointing task and four experimental conditions, demonstrating that the simulated fatigue data matches human fatigue data. We also compare two effort models: 1) instantaneous joint torques commonly used in computer animation and robotics, and 2) the recent Three Compartment Controller (3CC-) model from biomechanical literature. 3CC- yields movements that are both more efficient and relaxed, whereas with instantaneous joint torques, the RL agent can easily generate movements that are quickly tiring or only reach the targets slowly and inaccurately. Our work demonstrates that deep RL combined with the 3CC- provides a viable tool for predicting both interaction movements and user experiencein silico, without users.
Original languageEnglish
Title of host publicationProceedings of the 2020 CHI Conference on Human Factors in Computing Systems
Place of PublicationNew York, NY, USA
PublisherAssociation for Computing Machinery (ACM)
Pages1–13
Number of pages13
ISBN (Print)9781450367080
DOIs
Publication statusPublished - 2020
MoE publication typeA4 Article in a conference publication
EventACM SIGCHI Annual Conference on Human Factors in Computing Systems - Honolulu, United States
Duration: 26 Apr 202030 Apr 2020
https://chi2020.acm.org/

Conference

ConferenceACM SIGCHI Annual Conference on Human Factors in Computing Systems
Abbreviated titleACM CHI
CountryUnited States
CityHonolulu
Period26/04/202030/04/2020
Internet address

Keywords

  • biomechanical simulation
  • reinforcement learning
  • computational interaction
  • user modeling

Fingerprint Dive into the research topics of 'Predicting Mid-Air Interaction Movements and Fatigue Using Deep Reinforcement Learning'. Together they form a unique fingerprint.

Cite this