To enable a more natural feeling prosthetic hand interface, researchers from Florida Atlantic University’s College of Engineering and Computer Science and collaborators incorporated stretchable tactile sensors using liquid metal on the fingertips of a prosthetic hand. Encapsulated within silicone-based elastomers, the technology provides key advantages over traditional sensors, including high conductivity, compliance, flexibility, and stretchability. For the study, published in the journal Sensors, researchers used individual fingertips on the prosthesis to distinguish between different speeds of a sliding motion along different textured surfaces. The four textures had one variable parameter: the distance between the ridges. To detect the textures and speeds, researchers trained four machine learning algorithms. For each of the ten surfaces, 20 trials were collected to test the ability of the machine learning algorithms to distinguish between the ten different complex surfaces comprised of randomly generated permutations of four different textures.
Results showed that the integration of tactile information from liquid metal sensors on four prosthetic hand fingertips simultaneously distinguished between complex, multitextured surfaces—demonstrating a new form of hierarchical intelligence. The machine learning algorithms were able to distinguish between all the speeds with each finger with high accuracy. The technology could improve the control of prosthetic hands and provide haptic feedback.
“Significant research has been done on tactile sensors for artificial hands, but there is still a need for advances in lightweight, low-cost, robust multimodal tactile sensors,” said Erik Engeberg, PhD, the study’s senior author. “The tactile information from all the individual fingertips in our study provided the foundation for a higher hand-level of perception enabling the distinction between ten complex, multitextured surfaces that would not have been possible using purely local information from an individual fingertip. We believe that these tactile details could be useful in the future to afford a more realistic experience for prosthetic hand users through an advanced haptic display, which could enrich the amputee-prosthesis interface and prevent amputees from abandoning their prosthetic hand.”
Researchers compared four different machine learning algorithms for their successful classification capabilities: K-nearest neighbor (KNN), support vector machine (SVM), random forest (RF), and neural network (NN). The time-frequency features of the liquid metal sensors were extracted to train and test the machine learning algorithms. The NN generally performed the best at the speed and texture detection with a single finger and had a 99.2 percent accuracy to distinguish between ten different multitextured surfaces using four liquid metal sensors from four fingers simultaneously.
Editor’s note: This story was adapted from materials provided by Florida Atlantic University.
Photograph: Researchers used individual fingertips fitted with stretchable tactile sensors with liquid metal on a prosthesis attached to a robotic arm. Photograph by Alex Dolce courtesy of Florida Atlantic University.