Multimodal Interfaces for Upper-Limb Rehabilitation Robotic Machines

Stroke is one of the leading cause of permanent disability. Numerous rehabilitation methodologies can be adopted to deal with stroke motor impairment, but the optimal training approach remains unclear. Robotic technologies are recognized as powerful tools to promote neuroplasticity and stimulate motor re-learning. They allow delivering high-intensity, repetitive, active and task-oriented training; in addition, they provide objective measurements for patient evaluation. However, typical stroke robot-aided rehabilitation is very task specific and conceived for applying a sort of "if-then" algorithm: if the patient performs an incorrect motion, then the robot applies a predefined unidirectional action. Nowadays, the increasing employment of bio-cooperative systems for the administration of robot-aided therapy to stroke patients represents a real breakthrough. Such systems place the patient in the control loop, by feeding back his/her biomechanical and physiological state, and automatically adapt the control system on the basis of the monitored patient states, with the aim of motivating in an engaging and challenging way the patient. This work presents the design and the development of a bio-cooperative system for upper-limb robot-aided therapy consisting of two main modules: (i) a multimodal interface, able to evaluate patient performance and realtime display a virtual reality environment relative to the performed task, (ii) an adaptive robotic controller, able to modulate the complexity of the assigned motor task, allowing the administration of a patient-tailored therapy. A special attention has been paid to the development of an unobtrusive multimodal interface for monitoring patient state without altering user natural motion. Furthermore, a reliable method for reconstructing the 7-DoF upper-limb kinematics during robot-aided tasks with end-effector machines is presented and experimentally validated. Finally, two applications of the developed bio-cooperative system are presented: (i) a robotic system for 3D upper-limb rehabilitation, (ii) the quantitative evaluation of the outcomes of a clinical study involving chronic stroke patients. Preliminary experimental results of both the applications are reported.