Development of a Combined Myoelectric and Near-infrared Sensor for Prostheses Control

The control of upper-limb prostheses is commonly realized via the detection, processing and classification of myoelectric signals. Pattern recognition techniques enable users to control prosthetic devices by contracting remaining muscles, so that up to nine hand positions can be distinguished with high accuracy. It is however difficult to identify single fingeror sophisticated finger-movements. This is due to the fact, that corresponding muscles produce crosstalk, entailing it impossible to distinguish delicate finger motion. For this reason, we propose sensor fusion incorporating a near-infrared spectroscopy sensor, to detect changes of oxygen concentration in the blood, and an electromyographic sensor. Experiments show, that the sensor combination allows to distinguish between contracted and non contracted muscles. In this contribution, we present a miniature combined myoelectric and near-infrared sensor that can be used for prostheses control. The sensor uses LEDs instead of Lasers, which ensures safe operation, cheap production and consumes little power. The nearinfrared signal can be used as additional classifier input. We fount, that the spatial resolution of the combined sensor is better than common electromyographic sensors.