Vision-Based Yaw Bias Estimation for a Quadrotor UAV in Indoor Environment

Indoor UAV applications have attached a great attention currently provided that a series of high localization precision technology systems are laid out. UWB technology has been widely used in indoor UAV flying for its low cost and decimeter-level positioning. For successful flight, it is essential to determine the bias between the yaw measurement from (Attitude and heading reference system) AHRS onboard UAV and the true value of yaw angle in the UWB positioning system. Considering the low quality of the AHRS, we develop a method for estimating the yaw bias when a marker imaged from a small UAV. Classically this has to rely on a user to calibrate the AHRS manually before the UAV takes off. In this work multiple bearing measurements of a marker taken are employed to eliminate the yaw bias automatically when the UAV flies. The main result of this paper is that the yaw bias can be accurately estimated in real time based on these bearing measurements. The efficiency of this technique is validated by simulation results and flight test.