Control Tuning for the Quadcopter Unmanned Aerial Vehicle Based on Genetic Evolutionary Algorithm

Abstract

The quadcopter is an unmanned aerial vehicle that has been used in civil and military applications because of its relatively low cost, simple design, and ease of maintenance in addition to its unique flying characteristics and vast potential. Six Proportional-Integral-Derivative (PID) controllers are usually built for the position and attitude control of the quadcopter. However, tuning the controllers’ parameters poses a real challenge, as the system is highly coupled and contains two cascaded control loops while being underactuated (only four control inputs for the six degrees of freedom). In this study, a quadcopter model has been implemented by using MATLAB / SIMULINK where the model has been constructed to be integrated with the evolutionary genetic algorithm in addition to a third-party toolbox that is used to simulate fractional order calculus. The gains of the traditional and the fractional-order proportional-integral-derivative controllers have been tuned by using the genetic algorithm against different path shapes. The performance of the manually and the genetic algorithm tuned proportional-integral-derivative controllers have been compared with the genetic algorithm tuned fractional-order proportional-integral-derivative controllers. The genetic algorithm tuned fractional-order proportional-integral-derivative controllers have been superior and more robust in all cases with a lower fitness function value, much smother responses, and higher trajectory tracking ability.