Three-dimensional attitude determination using vector observations and optical flow sensors

Abstract

Flight control of unmanned aerial vehicle relies on three-dimensional attitude determination, and the most commonly used solution is based on the combination of magnetometer, accelerometer, and gyroscope, namely the MARG sensor. Optical flow sensor is also widely applied on UAV to detect the motion of UAV relative to the ground. In this paper, we proposed a novel method that using multiple optical flow sensors to measure angular velocity. This method utilizes a linear model to resolve both linear and angular motion from the measurements of multiple optical flow sensors. In the case of in situ rotations, experiment results prove that the proposed method can effectively estimate the angular velocity, and it can be used to augment MARG-based attitude determination.