The application of extended Kalman filtering to autonomous formation flight of small UAV system

Purpose - The purpose of this paper is to present a small UAV system with autonomous formation flight capability, the Swallow UAV system, and the application of an extended Kalman filter (EKF) based augmentation method to reduce the impact of data link loss, which will fail the formation flight algorithm of the system. Design/methodology/approach - The hardware of the Swallow UAV system is composed of two aircraft and a set of ground control station for leader-wingman formation flight. A hardware-in-the-loop simulation environment is build to support the system development. Fuzzy logic control method is applied to the guidance, navigation, and control system of leader and wingman aircraft. The leader system is designed with waypoint navigation and circle trajectory tracking functions to make the aircraft stay in visual range autonomously for safety. The wingman system is designed with formation flight functionality. However, the relative position and velocity are derived from the wireless data link transmitted leader navigation information. It is vulnerable to the data link loss. The EKF based leader motion estimator (LME) is developed to estimates the leader position when the data link broke, and corrects the estimation when the data link is available. Findings - The designed LME is flight tested, and the results show that it woks properly with sound performance that the estimation error of relative position within 3 meters, relative velocity within 1.3 meters, and leader attitude within 1.6 degrees in standard deviation. Originality/value - The research implements the autonomous formation flight capability with the EKF based LME on a small UAV system.