INS5-0364
Programming Physical Anisotropy in Polymer Composites for Materials-Encoded Somatosensation in Soft Robotics
When and Where
Sep 29, 2026
12:05 - 12:20
Presenter(s)
Minjeong Ha (Gwangju Institute of Science and Technology (GIST))
Co-Author(s)
Abstract
Processing multimodal stimuli through centralized computation causes severe bottlenecks in robotics. To overcome this, we propose materials-encoded somatosensation, in which the material itself acts as a processor. By engineering magnetic field-programmed physical anisotropy in multiscale polymer composites, we embed artificial peripheral nerves directly into the material and achieve highly selective sensing and deterministic motor control without extensive computational overhead. First, we demonstrate printable proximity sensors based on a magnetic microflake-polymer composite paste, in which the engineered magnetic anisotropy of the microflakes induces magnetoresistance that transduces fields into selective electrical signals. This non-contact proximity sensing penetrates physical obstacles and supports robust on-skin gesture recognition and touchless human-machine interfaces. Next, we build a closed-loop system that integrates magnetic origami actuators with compliant magnetic sensors, where real-time feedback guides shape-morphing actuation and translates stimuli into autonomous, reconfigurable motion at the hardware level. Finally, we realize materials-encoded somatosensation in a monolithic magneto-piezoelectric polymer composite, whose field-programmed spatial anisotropy of piezoelectric nanofillers autonomously decouples vector-resolved force modalities. Ultimately, programmed physical anisotropy enables these polymer composites to process multimodal stimuli and direct motion. This approach establishes a transformative materials platform for next-generation neuromorphic embedded systems.
