Three Axis Aircraft Autopilot Control Using Genetic Algorithms : An Experimental Study

The behavior of an aircraft can be described with a set of non-linear differential equations by assuming six degrees of freedom (3 for linear motions and 3 for angular motions) about x, y & z axis. All the aircrafts have a PID controller for autopilot control system for pitch, yaw and roll. The PID [6, 12] controllers are associated with their PID gains which can either be tuned manually or by optimized methods like genetic algorithms to get better control and enhancement in the performance. Genetic algorithms (GAs) are the powerful tool in the field of global optimization to various problems. Genetic algorithms are the models of machine learning which are based on the process of natural evolution. Here genetic algorithms are successfully applied for three axis autopilot (Pitch, Roll and Yaw) control of an aircraft using MATLAB simulation for tuning of PID controller and thus optimized values of PID gains for the controller are derived. Further the autopilot system has been validated using FlightGear software [1]. In this software the PID gains evaluated using genetic algorithms have been fed as an input in the autopilot program for the Boeing 747 - 400 (the model used here for autopilot system is Boeing 747 - 400).