Logo image
Predicting Mid-Air Interaction Movements and Fatigue Using Deep Reinforcement Learning
Conference proceeding   Open access

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

Noshaba Cheema, Laura A. Frey-Law, Kourosh Naderi, Jaakko Lehtinen, Philipp Slusallek, Perttu Hamalainen and ACM
PROCEEDINGS OF THE 2020 CHI CONFERENCE ON HUMAN FACTORS IN COMPUTING SYSTEMS (CHI'20), pp.1-13
01/01/2020
DOI: 10.1145/3313831.3376701
url
https://doi.org/10.1145/3313831.3376701View
Published (Version of record) Open Access

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.
Computer Science Computer Science, Cybernetics Computer Science, Information Systems Computer Science, Interdisciplinary Applications Computer Science, Theory & Methods Science & Technology Technology

Details

Metrics

Logo image