Theses and Dissertations - UTRGV

Date of Award


Document Type


Degree Name

Master of Science (MS)


Mechanical Engineering

First Advisor

Dr. Tohid Sardarmehni

Second Advisor

Dr. Constantine Tarawneh

Third Advisor

Dr. Horacio Vasquez


This paper presents a quadrotor controller using reinforcement learning to generate near-optimal control signals. Two actor-critic algorithms are trained to control quadrotor dynamics. The dynamics are further simplified using small angle approximation. The actor-critic algorithm’s control policy is derived from Bellman’s equation providing a sufficient condition to optimality. Additionally, a smoother converter is implemented into the trajectory providing more reliable results. This paper provides derivations to the quadrotor’s dynamics and explains the control using the actor-critic algorithm. The results and simulations are compared to solutions from a commercial, optimal control solver, called DIDO.


Copyright 2023 Edgar Adrian Torres. All Rights Reserve.