Biomimetic hairy affective-touch sensory AI interface

Area of Science:

• Biomimetic engineering
• Neuroscience
• Artificial Intelligence

Background:

• Artificial intelligence (AI) lacks the ability to perceive tactile emotions, limiting its understanding of human social behaviors conveyed through physical touch.
• Current AI systems struggle to interpret the nuances of affective touch, hindering natural human-AI interaction.
• Developing tactile sensing capabilities is crucial for AI to decode social cues embedded in physical contact.

Purpose of the Study:

• To create a biomimetic hairy sensing interface capable of capturing the spatiotemporal and intrinsic features of affective touch.
• To enable accurate emotion recognition in AI by replicating biological tactile sensing mechanisms.
• To establish a bioelectronic analog for affective touch transduction.

Main Methods:

• Development of a biomimetic hairy sensing interface with a bistage hairy structure and homogeneous nanomesh manufacturing.
• Utilizing isoline theory for high force detection sensitivity and spatial precision.
• Direct induction of neuromimetic electric pulse signals without a spike coding circuit.
• Integration with a hybrid neural network (convolutional neural network and contextual large language model).

Main Results:

• The hairy interface achieved high force detection sensitivity (0.67 N⁻¹) and spatial precision (1.61 mm localization accuracy over 100 cm²).
• The interface successfully replicated the behavior of biological C-LTMRs (slowly adapting, Merkel nerve receptors).
• Real-time emotion recognition was achieved with 82.37% accuracy using the hybrid neural network, analyzing individualized touch patterns.

Conclusions:

• The developed neuromorphic tactile framework successfully enables emotion recognition from affective touch.
• This bioelectronic analog of affective touch transduction advances the development of AI systems capable of understanding human physical contact.
• The technology facilitates closed-loop human-AI emotional interaction, paving the way for humanoid robots with natural affective communication abilities.