Ant‐inspired Trimodal Hair‐Like BiVO 4 Memristor for Intelligent Diving‐Assistance Systems

ABSTRACT In biological systems, single‐modal stimuli are often insufficient for complex behaviors, whereas multimodal integration significantly enhances environmental perception and decision‐making. Inspired by ants, which synergistically process optical, chemical, and mechanical signals, the development of multimodal neuromorphic devices is highly desirable. However, most existing memristive systems remain limited to single electrical modulation, restricting the realization of integrated sensing‐memory‐computing platforms. In this work, an electro‐opto‐liquid trimodal memristor based on a hair‐like BiVO 4 architecture is developed, enabling the intrinsic integration of electrical synaptic plasticity, optical detection, and liquid sensing within single device. The antenna‐like hair‐like morphology promotes efficient multimodal signal coupling and interaction, resulting in stable conductance modulation, broadband photoresponse extending into the near‐infrared region, over 27,000 optical endurance cycles, and high sensitivity to liquid environments. Benefiting from these characteristics, the device demonstrates effective multimodal information fusion in neuromorphic computing, achieving a high underwater sensing and recognition accuracy of ∼92% with a low loss of ∼0.23. Furthermore, it enables dynamic visual recognition, attaining ∼91% accuracy in identifying the burst motion of fish schools. This work highlights a viable strategy for constructing compact, multifunctional neuromorphic systems toward complex environmental perception and adaptive intelligence.