Data Fusion for Displacement Estimation and Tracking of UAV Quadrotor in Dynamic Motion

Abstract

The fusion of MIMU and GPS data is generally used to estimate the displacement and tracking of quadrotor UAVs. Meanwhile, displacement estimation inaccuracies during dynamic motion often occur. This error is caused by noise and limited sensor sampling rate especially occurs when the quadrotor changes its attitude rapidly to generate an instantaneous horizontal force. This paper proposes data fusion based on Kalman filter to estimate orientation and displacement. Experiments were also carried out to verify displacement accuracy, i.e. in single-axis and multi-axis sensor motions. The algorithm combines data from MIMU and GPS sensors so that acceleration data is filled in points where GPS data is not available. With this method, the predicted displacement from the MIMU sensor can be corrected every second with data from the GPS and produce accurate displacement and trajectory estimates.