Mobile Robots for Dynamic Environments
3. Visual Attitude Estimation and Stabilization of Flying Robots
Download citation file:
- Ris (Zotero)
- Reference Manager
This chapter deals with visual attitude estimation in flying robots, and using this information to stabilize them. Unmanned Air Vehicles (UAVs) are currently in widespread use in many applications ranging from military operations to civilian tasks. Successful control of a UAV requires accurate and fast estimation of the vehicle attitude. Usually inertial navigation systems (INS) are used to obtain this attitude information in UAVs. As an alternative or as an additional sensor, vision systems can also be used to obtain vehicle states. Vision systems are readily available on various UAV platforms and can be used for this purpose. The use of vision for attitude estimation is reliable and affordable. In this chapter we present the use of a vision system that can be used to estimate vehicle attitude. Unlike previous works that use natural or artificial features such as blobs or parallel lines on the environment, this work involves the use of no special feature but the natural scene. Processing of this natural scene around the robot leads to attitude information which is used by the control algorithm to stabilize the robot. The vision processing and control are performed on board the vehicle using a vision computer. First, the algorithm, the UAV modeling and control are presented. A quadrotor flying robot is chosen as an experimental platform in this study. Second, a detailed presentation of the developed quadrotor system and experimental set-up are given. Finally, we present the experiments and the results of the estimation and control algorithms.