Optical lace offers soft robots a tactile sensor network
Their article “Optical lace for synthetic afferent neural networks” published in Science Robotics describes various experiments using a 3D-printer to integrate arbitrary 3D grids of soft stretchable light guides throughout the volume of a soft deformable scaffold. Some of the light guides are considered as input cores (receiving light from a LED) while networks of neighbouring light guides act as output cores, only carrying leaked light from the input cores upon deformation.
By optically measuring the coupling interactions (all the light power outputs), they were able to sense spatially continuous deformations and localize them with sub-millimetre accuracy, detecting forces as low as 0.3 Newton. In one experiment, they were able to simultaneously locate multiple finger presses and monitor the volumetric structural deformation of a soft scaffold just by measuring the coupling interactions within the optical lace. Here they used 1.5mm diameter polyurethane light guides for the input cores and 1mm diameter light guides for the output cores, loosely held in place into lattice-work channels designed within the scaffold structure and spaced apart by small air gaps in the resting state. For any given finger press, the touch position could be calculated using the ratio of intensities in the neighbouring output light guides.
Although the paper reports a minimum and maximum readable force of 1.5N and 5N respectively for the first OL scaffold experiment, the researchers note that the structure design could easily be tuned for different sensitivity and dynamic force ranges, using a different lattice geometry.
The lace itself, without being encapsulated in a lattice, had a minimum detectable force of 0.06N, but again, maybe it could be designed thinner and with a softer material. According to ballistic tests (shooting a small projectile at the innerved scaffold), measurable deformation rates go up to at least 46kHz with measurable impulses between 0.2 and 2.5ms.
The authors also anticipate they could further improve the positional accuracy of their tactile sensor network by changing the output geometries to be flatter and overlapping, with narrower light guides. Even a single input could innervate a large volume as long as higher power LEDs or more sensitive photodetectors are used.
Next, they want to take advantage of the higher information density that can be carried through optical systems to create integrated sensorimotor networks, combining not only deformation sensing but also temperature, humidity, and chemical monitoring.
Cornell University – www.cornell.edu