Human-inspired control strategy for hand prosthetics

The hand is the human body part that has always fascinated researchers: several studies have been conducted to understand and explain its perfect mechanism. The hand is used to learn and to interact with the environment and it is clear that hand loss involves irreparable damage for a person. Besides having suffered hand loss, monolateral amputee subjects need to learn how to perform everyday life actions with only one hand. To overcome this problem, since the ancient Egypt, prostheses have been used for both cosmetic and functional purposes. The functioning of an active prosthetic hand is guaranteed by a mechatronic device, a decoding system to decode human biological signals in gestures and a control law that translates all the inputs (from the hand and from the user) in the desired movement. The ambition of this thesis is to design and develop a control strategy able to improve grasp and manipulation capabilities of prosthesis based on the tactile sensorization and on the use of this information in the control strategy. The proposed strategy is divided into 3 levels: low-level to regulate force and slippage during the whole grasp, middle-level, to manage the pre-shaping and the fingers reaching to the object, high-level, to decode the biological human signals in movements and force levels. The greatest limitation for an amputee subject who uses a prosthesis having no sensory feedback is the difficulty to manage unexpected events in an autonomous way. In grasp and manipulation tasks, the possibility of the object slippage is high. For this reason, it is necessary to detect the beginning of the slippage and provide control with a fast contrast action. The first contribution of this thesis is the development of a touch-and-slippage detection algorithm for effective grasp control of a prosthetic hand embedding monoaxial, low-cost sensors is proposed. One of the main problems in the prosthetic hand design is to provide the hand with a reliable system for force and slippage control. To decrease the attention level and the cognitive burden for the user during grasp tasks, an automatic strategy is necessary. The second contribution of this thesis is to propose a force-and-slippage control strategy able to i) regulate the grasping force, ii) prevent the slippage events, iii) coordinate fingers for replicating a human-like behaviour on the prosthetic system. Real-time reaction to slippage events and finger coordination have been achieved by means of i) a force control with inner position loop, ii) a sensorization system giving information about the applied normal forces, and iii) an approach for controlling the fingers in a coordinated manner on the basis of the virtual finger concept. The middle-level is managed with a proportional position control where the user can actively close or open the hand. The high-level proposed in this thesis consists of a hierarchical classification system used to simultaneously discriminate hand/wrist gestures and desired force levels. Moreover, a system composed of software for the EMG signals management and virtual reality were ad-hoc developed for upper limb amputees underwent the Targeted Muscle Reinnervation (TMR) to train them to control multiple prosthetic modules in a coordinated manner in a safe environment. Dexterity and manipulations skills in humans are allowed by complex biomechanics of the hand and a control loop based on a bidirectional communication with the brain, thanks to a sophisticated sensory system. This thesis has contributed to show the possibility to enable real-time closed-loop control of bionic hands in tasks of fine grasp and manipulation. Force and slippage sensations were elicited in an amputee by means of biologically inspired slippage detection and encoding algorithms, supported by an extended stick-slip model of the performed grasp. Closed-loop control capabilities enabled by the neural feedback were compared with those achieved without feedback.