Potentials and limitations of RoboSkin: a MEMS pressure sensor based tactile sensing solution

Abstract

Tactile sensing has become a key enabling technology for robotic manipulation by providing robots with information about physical interactions, including contact, force, slip, and object properties. Among the various tactile sensing approaches, RoboSkin employs commercially available MEMS pressure sensors embedded within compliant elastomeric structures to realize compact, robust, and scalable tactile sensing systems. By combining low-cost hardware with customizable mechanical transduction layers, RoboSkin offers a versatile platform for applications ranging from dexterous manipulation to industrial gripping. This paper reviews the RoboSkin concept, covering its underlying measurement principle, manufacturing approach, and scalable system architecture. The evolution of the technology is discussed through developments in tactile sensor arrays, contact localization, force estimation, slip detection, multimodal sensing, and application-specific mechanical optimization. The reviewed implementations demonstrate that the modular architecture can be scaled from individual sensing elements to distributed arrays while retaining the same fundamental measurement principle. They further show that the geometry and material properties of the mechanical transduction layer strongly govern sensitivity, measurement range, spatial response, hysteresis, and cross-talk, making its design central to application-specific sensor performance. Finally, the key potential and current limitations of the RoboSkin concept are analyzed. Overall, RoboSkin has evolved into a versatile and scalable tactile sensing platform that provides a practical foundation for next-generation robotic manipulation systems.