Abstract
Flexible magnetic tactile sensors hold transformative potential in robotics and human-computer interactions by enabling precise force detection. However, existing sensors face challenges in balancing sensitivity, detection range, and structural adaptability for sensing force. This study proposed a pre-compressed magnetization method to address these limitations by amplifying the magnetoelastic effect through optimized magnetization direction distribution of the elastomer. A body-centered cubic lattice-structured magnetoelastomer featuring regular deformation under compression was fabricated via digital light processing (DLP) to validate this method. Finite element simulations and experimental analyses revealed that magnetizing the material under 60% compression strain optimized magnetization direction distribution, enhancing force-magnetic coupling. Integrating the magnetic elastomer with a hall sensor, the prepared tactile sensor demonstrated a low detection limit (1 mN), wide detection range (0.001-10 N), rapid response/recovery times (40 ms/50 ms), and durability (>1500 cycles). By using machine learning, the sensor enabled accurate 3D force prediction.