Control of Quadrotor Drone with Partial State Observation via Reinforcement Learning

In this paper, we propose a quadrotor control algorithm which use the historical strengthened partial state observations as input information to control the quadrotor drone using reinforcement learning algorithm. Reinforcement learning method could enable the agent to learn a policy which could map the observations to control commands, which, in our work, is the actuator command of the quadrotor. Besides, we conduct our method in the control task via simulation, the results of which show excellent performance.