Robust Adaptive Non-Linear Control Design for an Aerial Robot with In-Door Application in Constrained Corridors

Abstract

A robust adaptive non-linear controller for a quadcopter manipulator system, used in applications in constrained narrow corridors, is proposed in this paper. The aerial robot considered comprises a three degree of freedom manipulator attached to the aerial vehicle's center of gravity at the bottom. During tasks that involve constrained environments like corridors, the controller should provide efficient control inputs to traverse with minimum trajectory tracking error, disturbance rejection, and stability. External disturbances such as wind, noise, and other factors and unmodeled non-linearities within the model affect the system in real-time applications and degrade its performance. To achieve stability and minimize trajectory tracking error, a novel robust augmented adaptive torque control law is developed for the system, which combines a feedback linearization controller with a model reference adaptive controller. The integrated system has a coupled uncertain non-linear dynamics. The adaptive mechanism's update law is derived using the SPR-Lyapunov approach and then modified with a Γ-projection operator to ensure that the estimates are bounded. In addition, the designed controller with projection-based adaptive law is implemented on the unified plant dynamics and evaluated using MATLAB and ROS/Gazebo simulations. A real-time task scenario is developed in ROS/Gazebo, with two rooms connected by a small corridor as the environment and ArUco marks on the walls serving as targets.