The development of bioinspired switchable adhesive systems has promising solutions in various industrial/medical applications. Switchable and perceptive adhesion regardless of the shape or surface shape of the object is still challenging in dry and aquatic surroundings. We developed an electronic sensory soft adhesive device that recapitulates the attaching, mechanosensory, and decision-making capabilities of a soft adhesion actuator. The soft adhesion actuator of an artificial octopus sucker may precisely control its robust attachment against surfaces with various topologies in wet environments as well as a rapid detachment upon deflation. Carbon nanotube-based strain sensors are three-dimensionally coated onto the irregular surface of the artificial octopus sucker to mimic nerve-like functions of an octopus and identify objects via patterns of strain distribution. An integration with machine learning complements decision-making capabilities to predict the weight and center of gravity for samples with diverse shapes, sizes, and mechanical properties, and this function may be useful in turbid water or fragile environments, where it is difficult to utilize vision.
Keywords: actuator; biomimetics; carbon nanotube; strain sensor; wet adhesion.