Estimating Model Parameters and PID Tuning of a Quadrotor Using Real Measurement Data

The estimation, identification, and control process of autonomous aerial vehicles, is a subject of increasing interest because more and more applications such as surveillance, monitoring, etc., are depending on their existence. Flight simulations are executed with an experimental real-case quadcopter in an open test bench area. We obtain values for all roll, pitch, and yaw axis. This paper suggests a methodology to estimate a model that predicts the system's output by exploiting measurement data. The innovation introduced by the authors is the comparison of several linear models using the Matlab® System Identification Toolbox and the selection of the most flexible and the one that fits the operable behavior of the drone in a high percentage. For the control part of the paper, to better handle the UAV and all its variables, PID control techniques are proposed to enhance the behavior of the closed-loop system.