Electronic Skin: Opportunities and Challenges in Convergence with Machine Learning.

Recent advancements in soft electronic skin (e-skin) have led to the development of human-like devices that reproduce the skin's functions and physical attributes. These devices are being explored for applications in robotic prostheses as well as for collecting biopotentials for disease diagnosis and treatment, as exemplified by biomedical e-skins. More recently, machine learning (ML) has been utilized to enhance device control accuracy and data processing efficiency. The convergence of e-skin technologies with ML is promoting their translation into clinical practice, especially in healthcare. This review highlights the latest developments in ML-reinforced e-skin devices for robotic prostheses and biomedical instrumentations. We first describe technological breakthroughs in state-of-the-art e-skin devices, emphasizing technologies that achieve skin-like properties. We then introduce ML methods adopted for control optimization and pattern recognition, followed by practical applications that converge the two technologies. Lastly, we briefly discuss the challenges this interdisciplinary research encounters in its clinical and industrial transition.

High-Stretchability, Ultralow-Hysteresis ConductingPolymer Hydrogel Strain Sensors for Soft Machines.

Highly stretchable strain sensors based on conducting polymer hydrogel are rapidly emerging as a promising candidate toward diverse wearable skins and sensing devices for soft machines. However, due to the intrinsic limitations of low stretchability and large hysteresis, existing strain sensors cannot fully exploit their potential when used in wearable or robotic systems. Here, a conducting polymer hydrogel strain sensor exhibiting both ultimate strain (300%) and negligible hysteresis (<1.5%) is presented. This is achieved through a unique microphase semiseparated network design by compositing poly(3,4-ethylenedioxythiophene):polystyrene sulfonate (PEDOT:PSS) nanofibers with poly(vinyl alcohol) (PVA) and facile fabrication by combining 3D printing and successive freeze-thawing. The overall superior performances of the strain sensor including stretchability, linearity, cyclic stability, and robustness against mechanical twisting and pressing are systematically characterized. The integration and application of such strain sensor with electronic skins are further demonstrated to measure various physiological signals, identify hand gestures, enable a soft gripper for objection recognition, and remote control of an industrial robot. This work may offer both promising conducting polymer hydrogels with enhanced sensing functionalities and technical platforms toward stretchable electronic skins and intelligent robotic systems.

Beyond Human Touch Perception: An Adaptive Robotic Skin Based on Gallium Microgranules for Pressure Sensory Augmentation.

Robotic skin with human-skin-like sensing ability holds immense potential in various fields such as robotics, prosthetics, healthcare, and industries. To catch up with human skin, numerous studies are underway on pressure sensors integrated on robotic skin to improve the sensitivity and detection range. However, due to the trade-off between them, existing pressure sensors have achieved only a single aspect, either high sensitivity or wide bandwidth. Here, an adaptive robotic skin is proposed that has both high sensitivity and broad bandwidth with an augmented pressure sensing ability beyond the human skin. A key for the adaptive robotic skin is a tunable pressure sensor built with uniform gallium microgranules embedded in an elastomer, which provides large tuning of the sensitivity and the bandwidth, excellent sensor-to-sensor uniformity, and high reliability. Through the mode conversion based on the solid-liquid phase transition of gallium microgranules, the sensor provides 97% higher sensitivity (16.97 kPa-1 ) in the soft mode and 262.5% wider bandwidth (≈1.45 MPa) in the rigid mode compared to the human skin. Successful demonstration of the adaptive robotic skin verifies its capabilities in sensing a wide spectrum of pressures ranging from subtle blood pulsation to body weight, suggesting broad use for various applications.

Cutaneous Ionogel Mechanoreceptors for Soft Machines, Physiological Sensing, and Amputee Prostheses.

Touch sensing has a central role in robotic grasping and emerging human-machine interfaces for robot-assisted prosthetics. Although advancements in soft conductive polymers have promoted the creation of diverse pressure sensors, these sensors are difficult to be employed as touch skins for robotics and prostheses due to their limited sensitivity, narrow pressure range, and complex structure and fabrication process. Here, a highly sensitive and robust soft touch skin is presented with ultracapacitive sensing that combines ionic hydrogels with commercially available conductive fabrics. Prototypical designs of the capacitive sensors through facile manufacturing methods are introduced and a high sensitivity up to 1.5 kPa-1 (≈44 times higher than conventional parallel-plate capacitive counterparts), a broad pressure detection range of over four orders of magnitudes (≈35 Pa to 330 kPa), ultrahigh baseline of capacitance, fast response time (≈18 ms), and good repeatability are demonstrated. Ionogel skins composed of an array of cutaneous mechanoreceptors capable of monitoring various physiological signals and shape detection are further developed. The touch skin can be integrated within a soft bionic hand and provide an industrial robot and an amputee with robust tactile feedback when handling delicate objects, illustrating its potential applications in next-generation human-in-the-loop robotic systems with tactile sensing.

Surface Actuation and Sensing of a Tensegrity Structure Using Robotic Skins.

Tensegrity robots comprising solid rods connected by tensile cables are of interest due to their flexible and robust nature, which potentially makes them suitable for uneven and unpredictable environments where traditional robots often struggle. Much progress has been made toward attaining locomotion with tensegrity robots. However, measuring the shape of a dynamic tensegrity without the use of external hardware remains a challenge. Here we show how robotic skins may be attached around the exterior of a tensegrity structure, to both control and measure its shape from its surface. The robotic skins are planar, skin-like membranes with integrated actuators and sensors, which we use to transform a passive tensegrity structure into an active tensegrity robot that performs tasks such as locomotion. In addition, sensors placed on the ends of the tensegrity rods are used to directly measure orientation relative to the ground. The hardware and algorithms presented herein thus provide a platform for surface-driven actuation and intrinsic state estimation of tensegrity structures, which we hope will enable future tensegrity robots to execute precise closed-loop motions in real-world environments.