Operation Skill Learning for Quadcopter Based on Deep Learning Using Visual Information

Abstract

In recent years, the research on the unmanned control of the moving vehicle has been proceeded, and various robots and motor vehicles moving automatically begin to be applied to the society. However, the more complicated their environment is, the more difficult the autonomous vehicles move automatically. Even in such environment, sometimes an expert with the operation skill is able to perform the appropriate control of the moving vehicle. In this research, a method for human’s operation skill learning by using CNN (Convolutional Neural Network) setting visual information for input is proposed in order to learn more complicated environmental information. CNN is a kind of deep learning method, and it is known that it has high performance in the field of the image recognition. In this experiment, we also visualized the operation knowledge by using FNN (Fuzzy Neural Network) with obtained input-output maps to create fuzzy rules. To verify the effectiveness of this method, we conducted an experiment of operation skill acquisition by some subjects using the simulator of quadcopter.