Magnetically Programmable Liquid Metal Composites With Hardware‐Level Bionic Anisotropic Intelligence for Multiplexed Sensing

ABSTRACT

Real‐time decoding of multidimensional biomechanical forces is critical for clinical monitoring yet remains challenging. Most conventional flexible sensors are intrinsically isotropic and cannot distinguish directional force vectors. Consequently, they rely on computationally intensive backend algorithms to separate coupled signals, introducing latency and power consumption. Inspired by muscle spindles, we report a biomimetic sensor that achieves signal decoupling directly at the hardware level. We utilized a magnetic programming strategy to align high‐aspect‐ratio CoNi alloy wires within a liquid metal matrix. Density functional theory calculations reveal that oxygen‐mediated interfacial coupling stabilizes this heterostructure while maintaining electrical connectivity. The resulting magnetically ordered architecture exhibits a high mechanical‐to‐electrical anisotropy index of 4.0, with gauge factors reaching 12.77 in the parallel direction. The sensor maintains linearity across a strain range exceeding 200% and demonstrates robust structural integrity over 3000 fatigue cycles. Wearable demonstrations show that the device can physically decode risk‐associated motion patterns without complex algorithmic post‐processing. It maps knee valgus to a distinct temporal feature, T _delay (Cohen's d = −5.85), and differentiates plantar shear modes via a spatial anisotropic ratio. This hardware‐enabled feature separability establishes a paradigm for intrinsic multiplexed sensing, enabling immediate, low‐latency wearable diagnostics.