Modeling and control of the robotic arm equipping an AGV feeder in the primary aluminum industry

The subject of this thesis is the modeling and control of this 3 degree of freedom (DOF) robot arm equipping an AGV that must feeds some cells with the aluminum fluorine transported by the vehicle. The modeling part deals with the derivation of forward kinematics, inverse kinematics and dynamics of the robot arm, while in the control part will be developed various implementations in order to solve the problem of the mass varying during the task. The first approach used is a Proportional Integral with gravity compensation controller with the empty robot arm in order to have a controller that maintain the home configuration. The next one used, is the Inverse dynamics control that is used to give a reference benchmark for the future controllers. This last controller is used for the tracking of a desired trajectory that simulate the movements of the real manipulator. Sliding mode control approach is used for dealing with the problem of uncertainties and noise that are introduced by the varying mass of the manipulator. Other comparisons are made with sliding mode controllers with time delay estimation of the dynamics. The robustness and efficiency of these approaches are tested with the same polynomial trajectories in joint space. All the methods produce acceptable results, while the best one is the Non linear terminal sliding mode control with time delay estimation, that presents the smallest position error and reduced chattering in the torques thanks to its simplicity and model-free nature.