Learning to Control a Quadcopter Qualitatively

摘要: Qualitative modeling allows autonomous agents to learn comprehensible control models, formulated in a way that is close human intuition. By abstracting away certain numerical information, qualitative models can provide better insights into operating principles of dynamic system comparison traditional models. We show learned from traces, contain enough information allow motion planning and path following. demonstrate our methods on the task flying quadcopter. A model through motor babbling. Training significantly faster than training times reported papers using reinforcement learning with similar quadcopter experiments. collision-free trajectory computed by means simulation, executed reactively while dynamically adapting characteristics system. Experiments have been conducted assessed V-REP robotic simulator.