Localização, modelagem e controle de um helimodelo em ambientes internos

Abstract

This work addresses the robotization problem of a small-scale commercial helicopter. The problem solution involves three subtasks: design and implementation of algorithms for localization, development of the helicopter mathematical model and design and implementation of control laws. For localization, the combination of a stereo vision system and inertial sensors information via Kalman Filtering is proposed. The estimated states, which are linear and angular positions and velocities, along with the pilot commands are used to estimate, through system identification methods, the parameters of a gray box model. Once the helicopter model is identified, the controllers are designed. The first tests were performed in a platform that restricted the helicopter movements to variations in the attitude angles. In the following tests, the helicopter was controlled without any device to restrict its movements, using traditional PID cascade control. Experimental results show stabilization of all the helicopter degrees of freedom.