MATLAB/Simulink and ROS/Gazebo Co-Simulation of an Upper Limb Rehabilitation Robot for Kinematics and Dynamic Modeling

This paper presents a co-simulation platform of a 6-DOF rehabilitation robot for assessing kinematics algorithms and controller quality. The co-simulation platform includes MATLAB/Simulink and Gazebo on ROS framework. The robot's digital model is built in Gazebo using a 3D model from SolidWorks. Kinematics are computed in MATLAB/Simulink using Denavit-Hartenberg (DH) for forward, while inverse uses Jacobian with Damped Least Squares (DLS) combined with Gradient Projection (GP). The robot is modeled using the Euler-Lagrange method. A Sliding Mode Controller (SMC) is designed in MATLAB/Simulink to enhance motion control performance. The co-simulation environment enables data exchange between MATLAB/Simulink and Gazebo, optimizing computation and physical environment simulation. The controller's stability is proven using the Lyapunov theory. Finally, several simulation scenarios are conducted to evaluate the co-simulation environment's performance by comparing it with Simscape Multibody and assessing the controller performance.