Survey on Attitude and Heading Reference Systems for Remotely Piloted Aircraft Systems

Abstract

Attitude estimation is a critical task for the safe navigation of RPAS (Remotely Piloted Aircraft System). When no 3D positioning information is available, e.g. due to GNSS (Global Navigation Satellite System) blockage, the autopilot must use data from the rest of the onboard navigation sensors to fly the RPAS in a stable manner and reduce potential damage on the ground. Hence, using accurate and efficient Attitude and Heading Reference System (AHRS) algorithms is of vital importance for the safety of the system. Different AHRS algorithms based on Kalman Filtering (KF) or Extended Kalman Filtering (EKF) have been proposed in the literature but there are only few works comparing them using experimental data collected with real sensors. In this paper three different AHRS algorithms have been compared using real sensor data collected with a commercial system (Xsens MTi-G). Additionally, one of these algorithms has been implemented in two autopilot platforms developed by CATEC and their performances have been compared with the Xsens MTi-G system. For assessing the accuracy of AHRS systems, the attitude estimations provided by the system needs to be compared with the attitude values that are considered to be a ground truth. In this paper, a multi-camera based motion capture system from VICON Motion System Ltd has been used to obtain the attitude ground truth. This system can estimate the position of each marker with sub-millimetric accuracy and the attitude of the object with an accuracy of less than a tenth of a degree.