Experiments on 2-DOF Helicopter Using Approximate Dynamic Programming

A. Birge || A. Fandel || Dr. S. Miah


In recent years, the use of unmanned aerial vehicles has seen significant growth in commercial and military sectors. Motion control of such vehicles remains difficult due to precessional torques causing erratic movements. Conventional linear control techniques are widely used. This project aims to experiment with the use of an adaptive motion control strategy. Specifically, a model-based reinforcement learning strategy, approximate dynamic programming, was examined for use on a 2 degree-of-freedom (2-DOF) helicopter, the Quanser AERO.

The adaptive motion control strategy for a 2-DOF helicopter was implemented following electrical engineering methodologies. The proposed strategy was modified for the specific system which required modeling and mathematical analysis. Simulations and experiments were conducted in certain operating conditions. The motion control strategy was implemented to the Quanser AERO using Simulink, a Raspberry Pi, and a Raspberry Pi/smart phone. With successful implementation, a proof of concept can be attained for use in other applications using embedded systems.