5 Degree Of Freedom Upper Limbs Rehabilitation Robot Centroid Modeling And Neural Network Control

As an important branch of medical robot, the upper limbs rehabilitative robot is based on the theory of hemiplegic upper limbs relearning, it is a hot research topic both in and abroad and combines many technologies such as rehabilitation medicine, human anatomy, mechanics, computer science, and robotics, etc. It plays a huge role to improve the career of disabled persons and the quality of life of disabled persons. The subjects is a research of a kind of exoskeleton robot for upper limb rehabilitation whose joint correspond to those of the human body. It can simulate the rule of arm movement through fix the upper limbs with the patient to rehabilitate upper limbs with upper limb rehabilitation robot drive patients do exercise.The rehabilitation robot we study is a 5-dof upper limbs rehabilitation robot which was designed by human anatomy and human body size parameter. It’s very important to establish a precise dynamics model during study the 5-dof rehabilitation robot, firstly, we introduce some basic knowledge of rehabilitation robot kinematics in detail, such as coordinate transformation theory, the D-H method, the Jacobi matrix, established the kinematics function and analysis the relationship between velocity and the terminal position of rehabilitation robot. Secondly, we established the dynamics function of rehabilitation robot through lagrange method when simplified the model to a 3-dof upper arm model. We also acknowledge that use the center of mass theory take the place of rotational inertia theory to calculate the dynamics function after study the process of establishing dynamics function in detail. Finally, we can prove that the dynamics function established by center of mass theory and rotational inertia theory is true through simulink, it can be used in the research of robot according to the conclusion.At last, in order to eliminate the influence of random swing to the rehabilitation robot dynamics function during rehabilitation training process, we use the RBF neural network controller in the system of rehabilitation robot, and select a best fit RBF neural network according to the uncertain function. It can be proved effectiveness based on the simulink with this control method in matlab.