3D printed strain sensors for soft sensing and actuation

Abstract

A key missing technology for the emerging field of soft robotics is the provision of highly selective multidirectional and stretchable tactile sensing that can be easily integrated into a robot using simple fabrication techniques. Conventional strain sensors, such as strain gauges, are typically designed to respond to strain in a single direction and are attached on the external surface of a structure. In this PhD research project, direction-selective sensors have been developed based on constriction-resistive and microcracking mechanisms and 3D printing methods have been employed for integrating the sensors directly into/on soft robots. Using a carbon nanotube reinforced polylactic acid (PLA-CNT), both sensing elements and conductive interconnects are 3D printed. For the constriction-resistive sensors, the sensitivity and anisotropy can be adjusted by controlling the air gap between printed adjacent tracks, infill density, and build orientation relative to the main loading direction. In particular, sensors printed with a near-zero air gap, i.e., adjacent tracks form a kissing bond, can achieve a gauge factor of ~2300 perpendicular to the raster orientation and a gauge factor of ~1 parallel to the raster orientation. The maximum directional selectivity of this ultra-sensitive sensor is 50.5, which is unprecedented among multidirectional sensors so far. The high sensitivity stems from the progressive opening and closing of the kissing-bond between adjacent tracks. This sensor proved to be able to sense the tiny strain resulted from the propagation of the ultrasonic wave in a solid plate as well. The constriction-resistive strain sensors only can operate in a small strain range. To detect strains in large strain ranges (>50%), a simple, low-cost, and scalable method of printing PEDOT:PSS thin film strain sensor onto 3D printed TPU is introduced to create highly stretchable integrated piezoresistive strain sensors and stretchable conductors for soft actuators. High strain sensitivity of ~ 417 is achieved with a linear working strain range of up to 100% strain. The high sensitivity stems from the non-continuous fragmentation of the PEDOT:PSS sensing layer on the patterned 3D printed TPU substrate. Furthermore, by changing the printing orientation of the TPU substrate from 0 to 90 degrees, the PEDOT:PSS layer turned insensitive to strain i.e. became a good conductor. The perpendicular-oriented transducer is used as a conductor while the parallel-oriented transducer is a sensor. To demonstrate the impact of this technology, we fabricate a surgical soft tentacle gripper (SSTG) that is controlled using a sensorized glove. We demonstrate precision control of the catheter bending motion with high accuracy of 99%, which shows the potential of using our sensor technology in minimally invasive soft robotic surgeries.