Online simulation of mechatronic neural interface systems: Two case-studies

Abstract

Neural interface systems (NIS) are widely used in rehabilitation and prosthetics. These systems usually involve robots, such as robotic exoskeletons or mechatronic arms, as terminal devices. We propose a methodology to assess the feasibility of implementing these kind of neural interfaces by means of an online kinematic simulation of the robot. It allows the researcher or developer to make tests and improve the design of the mechatronic devices when they have not been built yet or are not available. Moreover, it may be used in biofeedback applications for rehabilitation. The simulation makes use of the CAD model of the robot, its Denavit-Hartenberg parameters, and biosignals recorded from a human being. The proposed methodology was tested using surface electromyography (sEMG) signals from the upper limb of a 25-year-old subject to control a kinematic simulation of a KUKA KR6 robot. It was also used in the design process of an actual lower limb rehabilitation system being developed in our laboratories. The 3D computational simulation of this robot was successfully controlled by means of sEMG signals acquired from the lower limb of a 26-year-old healthy subject. Both real-time and prerecorded signals were used. The tests provided researchers feedback in the design process, looking forward to new iterations in the detailed design and construction phases of the project.