Robust Adaptive Feedback Linearization Controller for an Aerial Robot Working in Narrow Corridor and In-door Environments

Abstract

A Feedback linearization controller-based robust adaptive controller is designed for an aerial robot working in a cluttered environment in this paper. Aerial robots are robots with floating bases that can reach difficult to get to locations, and the manipulator can perform various tasks. While operating in in-door environments and in narrow corridors, these systems will be affected by external disturbances which can bring the system close to walls or other objects in the environment. Moreover, the internal unmodeled system non-linearities adds further challenge to the design requiring the controller to be robust. In this work, a robust adaptive augmented torque control law is proposed and implemented on an aerial robot consisting of a quadcopter and a three degree of freedom manipulator. The feedback linearization controller can cancel the non-linearities present in the plant model and yield error dynamics. A model reference adaptive controller is incorporated to estimate the unknown, uncertain system parameters for the feedback linearization controller. The robust update law for the adaptive mechanism is obtained using the SPR-Lyapunov method and modified using the $\Gamma$ projection operator augments robustness to the controller. The proposed controller is implemented on the unified dynamics of the quadcopter manipulator system in MATLAB and ROS/Gazebo platforms to validate the design. In ROS/Gazebo, the aerial robot is simulated using an iris drone with a three degree of freedom arm and an indoor space consisting of two rooms connected by a corridor and filled with objects. The obtained performance results in ROS/Gazebo show the proposed controller’s efficacy.