Optimal control of robotic hand for rehabilitation using fractional order systems and EEG signal processing

Abstract

Technology development has made it possible to process brain signals and make decisions based on them. In this case, it is possible to provide the movement of a robot based on brain signals and the design of a suitable controller. This paper presents a control method based on cognitive robotics for mechanical hand movement. In this paper, firstly, the robotic hand is modelled using the fractional order theory, and the mathematical model of the robotic hand is obtained. The user's hand is then placed in different positions, and EEG signals are collected from the user; then, the signals are labelled based on the video and hand movements. Then we recognize and classify different modes of user hand movement using the designed algorithm, including preprocessing, use of Large Laplacian Filter, windowing, feature selection and extraction, and the algorithm used for classification. After that, based on the recognized class and the optimal controller design, the movement of the robotic arm is performed. This paper considers two classes for hand movement; we used fractional order optimal control to design the controller, and we solved the fractional order optimal control by a numerical-analytical method based on the Hamiltonian function and orthogonal polynomials. Finally, we propose a simulation and conclusion.