DEVELOPMENT AND CONTROL OF A SHOULDER JOINT FOR HUMANOID ROBOTICS APPLICATION

Abstract

Humanoid robotics represents an important sub-field of Robotics that has many applications, such as house assistance, surveillance, medical assistance, dangerous environments, repetitive works, and rehabilitation. Human joints are difficult to replicate in a humanoid robot. The available biological models may differ from the real-life robotic models. This work aims at modeling, building, and controlling a low-inertia, high-stiffness, tendon driven shoulder joint. Force and position sensors are integrated for further study of the system dynamics. A Python-V-REP model of the joint is developed to study the joint motion and to optimize its design. Also, a Python interface is used to visualize the control mechanism and to represent the force sensor and magnetic encoder values that facilitate monitoring and controlling the actuators’ status. We explored the possibility to use EEG (Electroencephalogram) signals to control the robotic joint, taking into account possible applications in telerobotics and as prosthesis. A parallel manipulator with three curved limbs can be used in a humanoid robots as a shoulder and wrist joints. Describing the kinematics of this type of platform is a challenging process. This work tried also to examine two methods to find the direct and inverse kinematics of the parallel manipulator. The successive screw method is applied to one limb of the manipulator with the assumption that the other two legs will follow the first one. The geometric method describes a change in the height of the manipulator where pulleys are located.