Self-sensing actuators with programmable actuation performances for soft robots

Designing soft robots that are able to perceive unstructured, dynamic environments and their deformations has been a long-term goal. Previously reported self-sensing soft actuators were mostly constructed via integrating separate actuators and sensors. The actuation performances and the sensing reliability are affected owing to the unmatched materials and weak connections. Realizing a seamless integration of soft actuators and sensors remains a grand challenge. Here, we report a fabrication strategy to endow soft actuators with sensing capability and programmable actuation performances. The foam inside the actuator functions as actuator and sensor simultaneously, effectively addressing the conformability and connection reliability issues that existed in current self-sensing actuators. The actuators are lightweight (a decrease of 58% in weight), powerful (lifting a load of 433 times of its own weight), and versatile (coupling twisting and contraction motions). Furthermore, the actuators are able to detect multiple physical stimuli with high reliability, demonstrating their exteroception and proprioception capability. Two self-sensing soft robotic prototypes, including a bionic bicep and a bionic neck, are constructed to illustrate their multifunctionality. Our study opens up new possibilities for the design of soft actuators and has promising potential in a variety of applications, ranging from human-robot interaction, soft orthotics, to wearable robotics.