Comparação entre controladores PID com ajuste de ganhos com algoritmo genético para um veículo aéreo não tripulado de asa fixa

Abstract

Unmanned Aerial Vehicles (UAVs) are small aircrafts capable of being controlled remotely or operate autonomously, without a human aboard. Due to technological advancements and its increasing use in civilian tasks, new models and new control techniques to increase the cost-benefit and its performance are emerging. In this context, this work is focused on fixed wing UAV - a low cost and high performance model, which flight autonomy is achieved with control strategies implemented in its autopilot unit. In this scenario, this work aims to compare two control strategies: PID for position and PID for position and velocity, to evaluate Cessna Skyhawk's performance on both of them since their complexity are different. This aircraft is available on X-Plane 11 and the control codes are implemented on M. Thus, the Software in the Loop (SIL) experiments are executed with the flight simulator combined with MATLAB. To accomplish the communication between both softwares, it was implemented a protocol based on User Datagram Protocol (UDP) to get real time values of angles, positions and velocities from the aircraft on the simulator. With these values both PID controllers can define the required effort to the UAV to achieve a stable flight. Those values returned from the controllers are commanded to the aircraft's control surfaces that are responsible for its control during the simulation. Through communication protocol implemented on both programs, reading and writing data can be performed on simulation real time.