Multi-modal prosthetic fingertip sensor with proximity, contact, and force localization capabilities

The lack of sensory feedback provided by prosthetic hands dramatically limits the utility of the device. Peripheral nerve interfaces are now able to produce stable somatosensory percepts for upper limb amputees. Sensors must be able to detect forces across the fingers of the prosthesis in a repeatable and reliable fashion. We solved this concern with a novel multi-modal tactile sensor which consists of an infrared proximity sensor and a barometric pressure sensor embedded in an elastomer layer with potential use in prosthetic devices. Signals from both sensors measure proximity (0-10 mm), contact (0 N), and force (0-50 N) and are combined to localize impact at five spatial locations and three angles of incidence. Here, we describe the sensor design, its characterization, and data analysis. We use Gaussian process regression to fuse the signals from both sensors to obtain calibrated force in Newton with an R-2 value of 0.99. We use supervised learning to localize probe position and direction with classification accuracies of 96% and 89%, respectively. The complementary nature of both sensors leads to several sensing modalities that no one sensor can provide on its own and the repeatable, reliable, and compact form of the sensor enables use in multi-functional prosthetic hands.