Key points are not available for this paper at this time.
Empowering intelligent robots with cognitive abilities that rival or even surpass those of humans is the key to enabling more complex and sophisticated interactions. Currently, most electronic skins (e-skins) perform excellently in distinguishing shallow object properties, but they still face significant challenges in perceiving deeper object properties. Here, inspired by the multisensory integrated cognition of echolocation and tactility in bats, a bionic bimodal active cognition (BBAC) e-skin is designed. This e-skin utilizes feature fusion machine learning methods to evolve from passive perception to advanced active cognition and, further, constructs a bimodal enhanced intelligent material cognition system based on multilayer perceptron. This system enables intelligent robotic hands integrated with BBAC e-skin to achieve synergistic cognition of dielectric constant, softness, and material types of 8 smooth surface objects through a simple proximity and contact action without strictly controlling the test conditions, significantly surpassing the ability of humans.
Wang et al. (Thu,) studied this question.