Comparison Method Q-Learning and SARSA for Simulation of Drone Controller using Reinforcement Learning

Nowadays, the advancement of drones is also factored in development a world surrounded by technologies. One aspects emphasized here difficulty controlling drone, and system developed still under full control users as well. Reinforcement Learning used to enable operate automatically, thus drone will learn next movement based on interaction between agent environment. Through this study, Q-Learning State-Action-Reward-State-Action (SARSA) are study comparison results involving both performance effectiveness simulation methods can be seen through analysis. A Q-learning systems autonomous application was performed for evaluation study. According process shows that better effective train achieve desire compared with SARSA algorithm controller.