Similarity-Based Adaptive Complementary Filter for IMU Fusion

This paper addresses the attitude estimation problem using vector and gyroscope measurements. We propose a novel adaptation scheme for the complementary filter cut-off frequency which is based on the similarity between independent estimates obtained from the vector and gyroscope measurements. The adaptive complementary filter is also derived on the special orthogonal group and convergence of the filter is established. The effectiveness of our approach is demonstrated with simulation results.