A QoE Evaluation of Haptic and Augmented Reality Gait Applications via Time and Frequency-Domain Electrodermal Activity (EDA) Analysis

Electrodermal Activity (EDA), also known as skin conductance, is related to the peripheral response arising from the activation of the sympathetic nervous system. Changes in EDA can indicate emotional conditions such as stress, excitement, and fear. The literature lacks a full understanding of actual EDA signals and what EDA responses mean for various types of stimuli. EDA is difficult to capture accurately in applications where a person is moving such a walking, exercising, running. Signal processing tools can be used to inform if a device is capturing actual EDA signals and not artifacts and noise resulting from movement. The work presented here is an extension of a Quality of Experience (QoE) evaluation of a gait feedback system. The work presents a QoE evaluation and capturing protocol of a gait feedback system composed of two modules: haptic and Augmented Reality (AR). These stimuli were presented to a user if they needed to change their gait and knee misalignment. The key contribution of this paper is a report and analysis of frequency domain EDA responses of participants to these AR and haptic feedback stimuli. Here we include the capturing and signal processing pipeline for EDA capture during user movements. Without such signal processing, the captured data can provide unreliable results. Analysis between groups report statistically significant differences in EDA based on the status of the participants gait (e.g., varus, partial alignment (only one leg aligned) and complete aligned (both legs aligned)) for haptic and AR groups in time and frequency domains. These findings showed that our processing pipeline could detect actual EDA signals during movement (walking). In terms of the gait feedback, there was an increase in EDA events for haptic groups when compared to AR.