| C. Troussas, A. Krouska, Ph. Mylonas, C. Sgouropoulou, I. Voyiatzis |
| FuzzyHCI: A Type-2 Fuzzy Logic Framework for Real-Time Adaptive Human-Computer Interaction Based on Cognitive Load and Interaction Fluency |
| 11th South-East Europe Design Automation, Computer Engineering, Computer Networks and Social Media Conference (SEEDA-CECNSM 2026), Thessaloniki, Greece, September 18-20, 2026 |
|
ABSTRACT
|
| Adaptive HCI attempts to change the behavior of an interface depending on the user's state, but currently systems used either static thresholds which are too crisp to capture the nature of uncertainty present in human cognitive and behavioral states. This paper introduces FuzzyHCI a framework that uses interval type-2 fuzzy logic to capture the uncertainty of a dynamically estimated user state and create fine-grained interface adaptations. The framework uses three inputs - estimated cognitive load (derived from user actions and physiology) , task performance (derived from interaction features), and user expertise - which is mapped using a type-2 fuzzy inference system of 36 rules to output adaptive recommendations on interface complexity, information density, assistance level, and navigation layout. FuzzyHCI does not have clear state transitions unlike the rigid binary and crisp state-based rule systems, hence ensuring smooth, fine-grained adaptation without "state-hopping" or "transition shock". We conducted an experimental study with 186 participants attempting five types of tasks at three different levels of expertise. We found that FuzzyHCI adapts the interface effectively, leading to a 13% increase in Task completion rate (89.1% vs. 81.3% vs. 72.2% for FuzzyHCI, crisp-based, static), a 9.4% improvement in System Usability Scale (81.7 vs. 74.1 vs. 67.5) and a 45.7% reduction in user's cognitive load (compared to 22.3% reduction by crisp rules, and 31.2% increase by static interface over 60 minutes). The advantage of FuzzyHCI over crisp rules is particularly significant with novice users and complex multi-task environments; in these situations, the smooth transition provided by FuzzyHCI prevented overwhelming the user.
|
| 18 September, 2026 |
| C. Troussas, A. Krouska, Ph. Mylonas, C. Sgouropoulou, I. Voyiatzis, "FuzzyHCI: A Type-2 Fuzzy Logic Framework for Real-Time Adaptive Human-Computer Interaction Based on Cognitive Load and Interaction Fluency", 11th South-East Europe Design Automation, Computer Engineering, Computer Networks and Social Media Conference (SEEDA-CECNSM 2026), Thessaloniki, Greece, September 18-20, 2026 |
[
BibTex] [
Print] [
 Back] |
