MULTIVARIABLE PID CONTROL OF A TWO-DEGREE-OF-FREEDOM HELICOPTER PROTOTYPE USING EVOLUTIONARY ALGORITHMS

Abstract

The development of helicopter control applications represents a great challenge due to its several degrees of freedom, its predominantly non-linear dynamics and the strong coupling between its variables. To tackle this problem, a scale prototype of a two-degrees-of-freedom (2-DOF) helicopter was built using low-cost materials and open source technologies (Arduino and Scilab / Xcos). The prototype model is obtained with the MATLAB systems identification toolbox and then the interaction between the system variables is analyzed. As a control strategy, ideal PID-2GdL controllers with first order filter and static decoupling were used. The controllers were tuned to the evolutionary algorithms: particle swarm optimization (PSO) and cuckoo search (CS), minimizing by batch simulation a multiobjective function that contemplated the performance indices IAE and IAU. The best models identified in open-loop with the prediction error method (PEM) presented Box Jenkins and ARMAX structures with a moderate precision, which yielded the minor errors during the estimation and validation. The applied control technique allows obtaining PID controllers with excellent performance, low variability, adequate robustness for reference tracking and load disturbance rejection with a minimum error in steady state and an acceptable control effort.