Bioinspired Sensory Transduction for Magnetic Profile Recognition and Encryption

In biological systems, adaptive responses to environmental stimuli are facilitated by sensory transduction, where receptors transform stimuli into dynamic intermediate electrical signals for further processing. For bioinspired artificial systems, this suggests the need to develop concepts that transduce various stimuli into electrical intermediates for recognition. Inspired by biological magnetoreception in elasmobranchs, which sense magnetic environmental profiles for navigation, we introduce an artificial sensory transduction system for magnetic profile recognition of objects using electromagnetic induction to generate electrical intermediate signaling, coupled with machine learning for decoding. We design moldable magnetic soft composites (MSCs) comprising magnetic particles in a zwitterionic polymer matrix, encoding with both static (shape, rheology, and magnetization) and dynamic (magnetization decay) multidimensional features. Upon translocation through a receiving coil, MSCs generate distinct transient induced electrical signals. Machine learning algorithms decode the static and dynamic information with ∼100% and 87.5% recognition accuracy, respectively, with a recognition strength of 3 bits and a large information-carrying capacity of 10⁶²-10⁹³⁴ possible encoded states. We suggest that electromagnetic induction in soft composites is a useful and generalizable concept for sensory transduction in emerging adaptive dissipative bioinspired materials, haptic systems, and soft robotics.