Interacting Multiple Model Navigation System for Quadrotor Micro Aerial Vehicles Subject to Rotor Drag

• Published in: Proceedings of the ... IEEE/RSJ International Conference on Intelligent Robots and Systems pp. 10705 - 10710
• Main Authors: Gomaa, Mahmoud A.K., De Silva, Oscar, Mann, George K.I., Gosine, Raymond G.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 24.10.2020
• Subjects: Drag; Filtering algorithms; Kalman filters; Mathematical model; Numerical models; Rotors; Vehicle dynamics
• ISSN: 2153-0866
• DOI: 10.1109/IROS45743.2020.9340796

This paper presents the design of an Interacting Multiple Model (IMM) filter for improved navigation performance of Micro Aerial Vehicles (MAVs). The paper considers a navigation system that incorporates rotor drag dynamics and proposes a strategy to overcome the sensitivity of the system to external wind disturbances. Two error state Kalman filters are incorporated in an IMM filtering framework. The first filter has a model that uses conventional Inertial Navigation System (INS) mechanization equations, while the second filter considers a dynamic model with rotor drag forces of the MAV. In order to support the two error state Kalman filters, the generic IMM algorithm [1] is modified for error state implementation, handle dissimilar state definitions, and adaptive switching during operation. Numerical simulations and experimental validation using the EuRoC dataset are conducted to evaluate the performance of the proposed IMM filter design for changing flight conditions and external wind disturbance scenarios.