MODELING AND INTELLIGENT CONTROL OF A DRONE

Abstract

This thesis tackles the modeling, design, and control of a Quadrotor unmanned aerial vehicle, with a focus on intelligent control and smart applications such as obstacle avoidance, robust trajectory tracking, visual soft landing, and disturbance compensation. It details the mathematical modeling opted for the simulation and the control. Furthermore, It describes the classic control methodology for both linear and nonlinear control techniques with interpreted simulations; The methodology is subsequently applied to develop an open-source autonomous quadrotor miniature model. In addition, advanced control theory has been applied using Adaptive Linear Quadratic Gaussian, Model predictive control, and intelligent Radial basis functions neural network for the robust tracking of generated trajectory for either obstacle avoidance or bio-inspired soft landing on a specially designed landing pad. The thesis depicts as well the adaptive optimal observation by an enhanced Kalman filter combined with Madgwick sensor’s data fuse. Control laws were mainly either mathematically derived or adaptively generated based on stability analysis using Lyapunov theory, The simulation incorporated several analytical comparisons to prove efficiency and compare the performance.