Extended Kalman filter based autonomous flying system for quadcopters

Abstract

This paper presents mathematical modeling, implementation and experimentation results of Extended Kalman filter (EKF) implemented on existing flight control algorithm which is used to control multi-rotor unmanned aerial vehicles such as quadcopters, hexacopters, and octocopters. Purpose of implementing the EKF is to improve flight performance and reliability of the vehicles during its autonomous navigation which may include automatic take-off landing, waypoint navigation, and to improve the robustness for wind disturbances at the same time. Initially vision positioning data were used as a ground truth to validate the EKF outputs. Then the filter is tested in real-time using a quadcopter and experimental results were presented and compared with raw sensor data to evaluate system performance.