Implementing a Low-Cost Vision-Based 3d Position-Controller on a Quad-Rotor

Indoor position control of a flying robot is one of the important and challenging issues in robotics, for no use of the commercial GPS systems is possible in such spaces. This paper provides a new method for position control of a flying robot with the help of the feedback from an on-board infrared camera and other motion sensors on the robot. One of the challenges is the implementation of the vision-based controller on the robot, which has a nonlinear dynamic and unstable behavior. The flight stability and positioning control sensory system includes an IMU with three axes accelerometers and three axes gyroscopes, a barometer and an infrared-detecting camera. The measured signals are transmitted through a wireless channel to a ground station. The navigation loop is closed through an image processing on the ground, which transmits the control commands back to the robot, such that the captured image from an infrared emitting target remains as close as possible to the center of the video frame. The behavior of the overall system is verified through experimental studies, which reveal the effectiveness of the implemented system.