Stick‐On Large‐Strain Sensors for Soft Robots

Abstract

DOI: 10.1002/admi.201900985 pneumatic and microfluidic structures,[1–8] dielectric elastomer actuators,[9–11] shape memory alloys,[12,13] responsive hydrogels,[14,15] and living cells.[16] Among them, soft pneumatic robots have attracted attention due to their complex motion, simple control input, and low-impedance interactions. For soft pneumatic robots, sensing techniques are important to improve accuracy and functionality when grasping or manipulating objects of different shapes and sizes. However, the limitations of existing sensing capabilities greatly restrict the applications for this type of soft robot. Traditional rigid sensing components on soft robots can limit the deformation and compliance of the underlying soft robotic structure. Existing strain sensors for soft robots mainly rely on stretchable electronic conductors, either in the form of 1) elastomeric conductors, which consist of elastomers embedded with conducting components such as silver nanowires, silver particles, carbon nanotubes, and graphene, or 2) liquid conductors, such as grease and liquid metal. The limitations of these sensors have been noted in previous studies. The conductivity of elastomeric conductors degrades due to contact defects between separated particles or liquid, especially under large deformation, as well as the resistance-strain hysteresis resulting from cyclic loading.[17] For liquid conductors like conductive paste, the localized plastic deformation resulting from ratcheting during cyclic loading accumulates and deteriorates conductivity as well.[18,19] The change in conductivity leads to signal drifts. In addition, liquid conductors are not biocompatible and need tight sealing to prevent oxidation and leakage, thus requiring additional technical effort. Soft actuators with integrated microchannels filled with conductive fluid have been fabricated via 3D printing.[20,21] But the intricate designs and complex manufacturing requirements dramatically increase the fabrication difficulty. It remains a major challenge to manufacture robust soft strain sensors for soft robots.[22–27] Compared to electronic conductors, ionically conductive hydrogels can be readily used as stretchable conductors,[28–30] and have recently enabled a new family of devices called hydrogel ionotronics.[31] Both the mechanical and electrical properties of hydrogels can be tuned on demand over a wide range. For example, hydrogels can be as soft as living tissues or as tough as natural rubber. As another example, the resistivity of hydrogels can vary from 18.2 MΩ m to 10−1 Ω m, depending on the type and concentration of salt.[32] In fact, hydrogels resemble ideal conductors, as their resistivity is a material Soft robots require sensors that are soft, stretchable, and conformable to preserve their adaptivity and safety. In this work, hydrogels are successfully applied as large-strain sensors for elastomeric structures such as soft robots. Following a simple surface preparation step based on silane chemistry, prefabricated sensors are strongly bonded to elastomers via a “stick-on” procedure. This method separates the construction of the soft robot’s structure and sensors, expanding the potential design space for soft robots that require integrated sensing. The adhesion strength is shown to exceed that of the hydrogel itself, and the sensor is characterized via quasi-static, fatigue, and dynamic response tests. The sensor exhibits exceptional electrical and mechanical properties: it can sense strains exceeding 400% without damage, maintain stable performance after 1500 loading cycles, and has a working bandwidth of at least 10 Hz, which is sufficient for rapidly-actuated soft robots. In addition, the hydrogel-based large-strain sensor is integrated into a soft pneumatic actuator, and the sensor effectively measures the actuator’s configuration while allowing it to freely deform. This work provides “stick-on” large-strain sensors for soft robots and will enable novel functionality for wearable robots, potentially serving as a “sensing skin” through stimuli-responsive hydrogels.