Upper limb amputation deprives individuals of their innate ability to manipulate objects. Such disability can be restored with a robotic prosthesis linked to the brain by a human-machine interface (HMI) capable of decoding voluntary intentions, and sending motor commands to the prosthesis. Clinical or research HMIs rely on the interpretation of electrophysiological signals recorded from the muscles. However, the quest for an HMI that allows for arbitrary and physiologically appropriate control of dexterous prostheses, is far from being completed. Here we propose a new HMI that aims to track the muscles contractions with implanted permanent magnets, by means of magnetic field sensors. We called this a myokinetic control interface. We present the concept, the features and a demonstration of a prototype which exploits six 3-axis sensors to localize four magnets implanted in a forearm mockup, for the control of a dexterous hand prosthesis. The system proved highly linear (R2 = 0.99) and precise (1% repeatability), yet exhibiting short computation delay (45 ms) and limited cross talk errors (10% the mean stroke of the magnets). Our results open up promising possibilities for amputees, demonstrating the viability of the myokinetic approach in implementing direct and simultaneous control over multiple digits of an artificial hand.

The replacement of a missing hand by a prosthesis is one of the most fascinating challenges in rehabilitation engineering. State of art prostheses are curtailed by the physical features of the hand, like poor functionality and excessive weight. Here we present a new multi-grasp hand aimed at overcoming such limitations. The SSSA-MyHand builds around a novel transmission mechanism that implements a semi-independent actuation of the abduction/adduction of the thumb and of the flexion/extension of the index, by means of a single actuator. Thus, with only three electric motors the hand is capable to perform most of the grasps and gestures useful in activities of daily living, akin commercial prostheses with up to six actuators, albeit it is as lightweight as conventional 1-Degrees of Freedom prostheses. The hand integrates position and force sensors and an embedded controller that implements automatic grasps and allows inter-operability with different human-machine interfaces. We present the requirements, the design rationale of the first prototype and the evaluation of its performance. The weight (478 g), force (31 N maximum force at the thumb fingertip) and speed of the hand (closing time: <370 ms), make this new design an interesting alternative to clinically available multi-grasp prostheses.

Partial hand amputation is the most frequent amputation level worldwide, accounting for approximately 90% of all upper limb amputations. Passive cosmetic prostheses represent one of the possible choices for its treatment, probably the most affordable one. However, these devices restore very limited motor function and subtle sensory feedback. The latter is an important component for restoring the body schema. In this work we present a simple yet potentially effective and low cost cosmetic digital prosthesis that embeds touch feedback; we dubbed this DESC-finger. It delivers short-lasting vibrotactile bursts when it makes and breaks contact with the environment, based on the Discrete Event-driven Sensory feedback Control (DESC) policy. One prototype was developed and used by one amputee at home, for two months. The effectiveness of the device was experimentally assessed by means of an interview and a virtual eggs test, which showed, albeit preliminarily, that time discrete feedback can improve the motor control of a partial hand prosthesis in daily life conditions. Besides targeting people that already use cosmetic digits, the DESC-finger targets those that do not use them complaining for loss of sensibility. The production costs and manufacturing process makes the DESC-finger suitable for exploitation in high- and low-income countries.

No abstract