Flexible three-dimensional (3D) tactile sensors powered by artificial intelligence (AI) have attracted growing attention recently, as they constitute the core hardware for facilitating human-machine interaction that integrates human intelligence (HI) and machine intelligence (MI). This technology not only bridges humans, machines and objects, but also promotes the deeper convergence of the physical world, the digital world and the biological systems. Compared with the mainstream tactile capture methods, such as the depth-of-field cameras, which are susceptible to ambient environmental interference, or silicon-based inertial sensors with rigid geometries incompatible with soft biological tissues, flexible tactile sensing hardware exhibits distinct advantages. These systems enable the accurate detection and recognition of 3D motion behaviors, including position, velocity, orientation, and gesture, thereby offering a more comprehensive and immersive integration experience. These capabilities align closely with the urgent demands for 3D-motion capture in emerging AI‑driven scenarios such as embodied intelligence, the metaverse, and humanoid robots. Therefore, we outline the recent advances in flexible 3D tactile sensing technologies, covering materials and devices, structures and theories, systems and applications. It also highlights key challenges, discusses potential solutions, and presents forward-looking perspectives for future development in this field.
Qi et al. (Sun,) studied this question.