Recognitive Tactile Sensor System Inspired by Human Skin Mechanoreceptors

The human skin is an astonishing sensory organ that provides extensive information about our surrounding environment. It can detect various stimuli such as pressure, vibration, texture, temperature, and even chemical substances. Taking inspiration from ion channels and receptors, which are key components of cutaneous sensory organs from a biological perspective, a new approach is being pursued to address the limitations of conventional touch sensors and enable machines to interact with the world in a more human-like manner. Specifically, skin sensory organs perceive tactile sensations through the simultaneous and complex reactions of diverse and abundant receptors. Each receptor possesses a unique working mechanism that selectively responds to specific stimuli. Unfortunately, prior research has overlooked the significance of this aspect and instead focused on independent studies or functions of individual receptors, neglecting to mimic structure of the skin or incorporate it adequately.<br/>The skin encompasses numerous transmission networks that detect stimuli and relay information to the brain. These networks consist of receptors that detect stimuli, ion channels that generate action potentials, and nerves that transport the electrical signals to the brain.Specifically, in the context of sensing physical stimuli, mechanoreceptors are generally classified into four types: Merkel discs (MD), Meissner corpuscles (MC), Ruffini endings (RE), and Pacinian corpuscles (PC). These mechanoreceptors play a crucial role in collecting information about the physical properties of objects we encounter in the environment, enabling us to perceive their shape, weight, texture, etc.<br/>To truly mimic human skin functionality, it is essential to design and manufacture a diverse array of sensors capable of obtaining various information such as pressure, sheer force, and tension from a limited size. This will allow for more accurate and diverse collection of mechanical stimuli from the surrounding environment. For instance, in our case, to emulate the vibrational touch sensing function observed in dermal papillae where MC is located, a bump-like structure and piezoelectric characteristics can be utilized for implementation. Additionally, to mimic the positional and functional features of piezo2, found in the MD, a cone-shaped ion gel and piezo gel can be employed. As for the RE, which has a dendritic branch form within the capsule and participates in stretch detection, this can be emulated using carbon nanotubes (CNTs) sheet with properties of tensile resistance within viscoelastic materials. Simulating the functions and features based on these receptor structures represents a future challenge that artificial skin should strive towards.<br/>This study aims to develop a tactile sensor system that recognizes objects by mimicking the mechanoreceptors present in human skin. The sensor system is designed by emulating the unique structures and functionalities of three types of receptors (MD, ME, RE) crucial for object recognition. These sensors are embedded in a material with similar elasticity to human skin. The integrated sensors are arranged in a way that separates receptors (stimulus detectors) and ion channels (signal generators), enhancing selectivity to specific stimuli. The manufactured sensors exhibit stretchability similar to human tactile organs and demonstrate trends comparable to biological touch experiments. Furthermore, to achieve effective object recognition, an AI-based signal processing algorithm is employed to analyze and learn from the data generated by the sensors. The material design for the object recognition sensor incorporates materials such as piezo gels and non-hydrogel ion gels, as well as substances like carbon nanotubes to control mechanical and electrical properties. The integrated sensor system successfully recognizes surface textures and various objects that are difficult to analyze using only one or two types of sensors.