کنفرانس مهندسی برق ایران

صفحه اصلی / سی و سومین کنفرانس بین المللی مهندسی برق

Classifying Human Spatial Navigation Anxiety Using Electrooculography Signals and Machine Learning Techniques

نویسندگان :

Saeed Mousavi¹ Sara Ashrafi² Mehdi Delrobaei³

1- Department of Electrical Engineering, K.N.Toosi University of Technology 2- Department of Mechanical Engineering, K.N.Toosi University of Technology 3- Department of Electrical Engineering, K.N.Toosi University of Technology

کلمات کلیدی :

Electrooculography،Spatial Anxiety،Spatial Navigation،Machine Learning

چکیده :

Spatial navigation, a vital cognitive function enabling orientation, route planning, and landmark recall, is negatively impacted by anxiety. In the present study, electrooculography (EOG) signals were employed to classify levels of spatial navigation anxiety. EOG data were recorded non-invasively and in real time from 27 participants during a controlled navigation task. Features related to blinks, saccades, and fixations were extracted and subsequently provided as inputs to k-nearest neighbors, support vector machine, and decision tree classifiers. These models were applied to categorize anxiety into two and three classes, achieving accuracies of up to 85.71\% and 74.43\%, respectively. Significant features, including fixation mean duration and concatenated saccade mean velocity, were identified as key indicators of anxiety. A negative correlation between spatial navigation anxiety scores and navigation performance was observed, confirming that higher anxiety levels diminish navigational abilities. The presented findings indicate that objective, real-time assessment of spatial navigation anxiety can be realized through EOG-based analysis combined with machine learning techniques, thereby facilitating improved monitoring and support in critical navigational environments.