Application of Enhanced Model Reference Adaptive Control (EMRAC) algorithms to the tracking problem of Space Robotic Manipulators

This master thesis proposes the application of the Enhanced Model Reference Adaptive Control (EMRAC) to a space robotic manipulator. The EMRAC aims to be an improvement of the more generic Model Reference Adaptive Control (MRAC) by adding an adaptive integral action and an adaptive switching term to compensate for slow and rapid changing disturbances. Further robustness is achieved by the adoption of the σ-modification strategy that prevents the drifting of the adaptive gains. The robotic arm will replicate a realistic space mission of rescuing a faulty non cooperative micro-satellite. To the best of the author's knowledge, this project application is a novel implementation of the EMRAC controller with σ-modification to a MIMO space robotic system. Simulations show that the EMRAC is capable of controlling highly nonlinear systems affected by parameter uncertainties and unmodelled dynamics, furthermore, the algorithm is robust against disturbances. The thesis presents experimental data from computer simulations of four different implementations of the EMRAC performed using ROS and Gazebo. From simulations data a set of performance indexes are computed, showing that the four EMRAC strategies outperform benchmark controllers equipped with a full-state feedback linearisation control action including two standard controllers (PD and PI strategies) and a robust controller.