Hybrid PID-Fuzzy controller for autonomous UAV stabilization

Abstract

In the last two decades, there have been great advances in using Unmanned Aerial Vehicles (UAVs) in several applications. This rapid growth is due to their ability to carry out several types of tasks with high flexibility aligned with reduced risks to human life and cost-effectiveness. The increasing demands for more complex assignments have improved UAVs' capabilities, propelling the rise of platforms with a high degree of autonomy for performing simultaneous tasks with less human intervention. One of the great challenges of maneuvering a UAV for a stabilized flight consists of a highly-coupled nonlinear system with fast dynamics. Thus, attitude control is an essential issue for stabilizing the vehicle or for keeping it in the desired orientation. The PID controller is often used due to its simple structure, good stability properties, and less dependence on the exact system model. However, its tuning process may become complicated in the face of the system nonlinearities. In this scenario, an adjustment error can cause it to lose its flight stability. Therefore, the main contribution of this work is the development of a Fuzzy-PID hybrid controller to control a quadrotor UAV's height stability. The work also considers a classic PID controller for comparison purposes with the proposed hybrid controller. The ROS/Gazebo platform was used to carry out the experiments with both control structures. From the results, it was possible to verify that both PID and Fuzzy-PID controllers could perform the attitude control of the UAV. However, the hybrid control strategy obtained some advantages, such as self-adjustment through system variations.