How is skin sensing for robots advancing?

Robots need numerous sensors and various sensing technologies to perform a variety of functions. However, for human-robot Interaction, the most important sensing could be performed in the skin — especially in those robots designed to be as lifelike as possible and physically interface with (touch) humans. To enhance the collaboration between humans and machines, the ultimate skin sensors could detect touch (pressure), temperature, distance, and other physical interaction parameters.

Touch sensors and would-be touch-sensing suppliers keep investigating an ideal skin-sensing technique. Recently, Stanford University researchers provided updated details of a touch-sensing project that was first reported in 2012. At that time, they announced that they had developed the first multi-layer self-healing synthetic electronic skin. Now they have the first demonstration of a multi-layer, thin film sensor that automatically realigns during healing. The researchers used polypropylene glycol (PPG) and polydimethylsiloxane (PDMS, a silicone polymer) for their touch sensor. Both materials have rubber-like electrical and mechanical properties as well as biocompatibility. They can be mixed with nano- or microparticles to enable electric conductivity.

The durable, multilayer polymer material softens and flows when heated and then solidifies when cooled with both materials designed to have similar viscous and elastic responses to external stress over the desired temperature range. When heated to just 70° C (158° F), self-alignment and healing can occur in about 24 hours instead of taking as long as a week at room temperature. By adding magnetic materials to the polymer layers, the synthetic skin not only healed but also could self-assemble from separate pieces.

Looking forward, the researchers plan to make the layers as thin as possible so they can stack layers of varying functions. In addition to the current prototype that was designed to sense pressure, additional layers to sense changes in temperature or strain could be included.

Testing touch sensors
One of the issues for viable touch sensing is testability. Research at the University of Louisville, associated with Louisville Automation & Robotics Research Institute (LARRI), a multi-disciplinary team of faculty, staff, and students, and the Next Gen Systems (NGS) Group of Kentucky Advanced Manufacturing Partnership for Enhanced Robotics and Structures (KAMPERS) has led to the development of an automated test bench that can be used to quickly test and characterize new skin sensors. The NGS Group conducts research on the Next Generation Microsystems that are increasingly small, cheap, integrated, and networked. Key aspects of the automated test bench include force feedback control to drive the test actuator, and computer vision functionality to guide alignment of the test actuator and sensors arranged in a 2D array.

Skin cells used in the research consisted of poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS), a polymer mixture that has the highest efficiency among conductive organic thermoelectric materials. The skin cells were developed as a strain gauge sensing element on linear interdigitated electrode (IDE) structures. A micro-patterned gold structure within the PEDOT:PSS was used as the strain element. With strain applied to the gold pattern, resistance in the conductive structure changes, so researchers can observe how skin cells differ from one to another based on how they were fabricated and the patterned structure deposited within the PEDOT:PSS.

Using the force controller, repeatable force tests were conducted to fairly examine the sensors in a cell. This established capability allows the testing and characterizing of sensors to compare the performance between the individual sensors, as well as sensors made using different manufacturing techniques or from different batches. With the automated test bench, researchers can determine which deposition techniques and geometry designs to pursue for future testing and development. The automated test bench is expected to be used to quickly test and characterize new skin sensors, as well as any type of internally developed strain-based sensors.

References
TechXplore, Layers of self-healing electronic skin realign autonomously when cut
Science (2023). Autonomous alignment and healing in multilayer soft electronics using immiscible dynamic polymers
Louisville Automation & Robotics Research Institute (LARRI)
EPSCoR KAMPERS
Goulet, Brian P., “Automatic testing of organic strain gauge tactile sensors.” (2022). Electronic Theses and Dissertations. Paper 3897.