Predicting Sensory and Affective Tactile Perception from Physical Parameters Obtained by Using a Biomimetic...

Area of Science:

• Materials Science
• Sensory Science
• Consumer Research

Background:

• Tactile perception influences product evaluation and consumer preferences.
• This perception operates hierarchically, involving sensory (e.g., soft, smooth) and affective (e.g., comfort, luxury) layers.
• Objective measurement of tactile properties is essential for product design.

Purpose of the Study:

• To predict sensory and affective tactile perception scores of materials.
• To utilize a biomimetic multimodal tactile sensor for objective tactile data acquisition.
• To evaluate the efficacy of regression models in predicting tactile perception.

Main Methods:

• A biomimetic multimodal tactile sensor was employed to measure physical parameters (force, vibration, temperature).
• 32 diverse materials (cosmetics, textiles, leather) underwent sensory and affective descriptor evaluations.
• Ten regression models were trained using sensor data to predict perception scores.

Main Results:

• The bagging regressor model achieved high prediction accuracy (R² > 0.6) for most sensory and affective descriptors.
• Specific descriptors like "moist" and "elastic" (sensory) and "pleasant" and "like" (affective) showed particularly high prediction accuracy.
• The sensor effectively captured physical parameters correlating with human tactile judgments.

Conclusions:

• Biomimetic tactile sensing provides a viable method for predicting human tactile perception.
• This approach supports efficient tactile design and product development across industries.
• Objective prediction of tactile qualities can enhance product innovation and consumer satisfaction.