Fuzzy cognitive map for visual servoing of flying robot

This study aims to use the image-based visual servoing (IBVS) method for smart control of a flying quadrotor robot. The fuzzy cognitive map (FCM) is a causal diagram which shows the relationship among the main components in the system. The relationships in an FCM model are determined by experts who have knowledge of system components and their governing relations. Since an FCM entails the main advantages of both fuzzy logic and neural networks, it is a suitable choice for designing intelligent vision-based systems. In this article, it is assumed that image characteristics lie within the field of vision of the camera. Maintaining image features within the field of vision of the camera is of particular importance in an underactuated quadrotor robot. The reason is that, to produce translation in the robot, the quadrotor roll and pitch angles must be changed. We solve this problem by limiting the input accordingly. First, certain features were presented via combining the perspective image moments. Subsequently, these features were deployed in the IBVS using FCM servoing. The obtained simulation results showed that, in spite of the existing challenges, the proposed method was implemented successfully to control the quadrotor robot.