A Starfish‐Inspired Hierarchical Multimodal Flexible Sensor for Decoupled Thumb Motion Perception and Virtual Human–Machine Interaction

ABSTRACT Wearable human–machine interfaces (HMIs) require sensing systems that combine high accuracy, mechanical compliance, and low system complexity. Here, we present a starfish‐inspired, hierarchically decoupled multimodal flexible sensor for high‐fidelity thumb motion perception and virtual human–machine interaction. Inspired by the distributed sensing mechanism of starfish tube feet, the architecture integrates radially distributed strain channels with spatially isolated pressure nodes. This design enables simultaneous sensing of global deformation and localized tactile stimuli while minimizing mechanical coupling. Finite element simulations and experimental characterization demonstrate effective strain amplification, low interchannel coupling, stable electromechanical performance, and long‐term durability. Combined with a Light Gradient Boosting Machine (LightGBM) classifier, the system achieves validation and test accuracies of 99.82% and 99.27%, respectively, for six thumb motions, with strong cross‐subject generalization. The decoded motions enable intuitive, real‐time control in a virtual environment without bulky sensor gloves or external tracking systems. This work establishes a bioinspired strategy for structurally decoupled multimodal flexible sensing and provides a promising platform for wearable HMIs and intelligent robotics.