Novel MARG-Sensor Orientation Estimation Algorithm Using Fast Kalman Filter

Abstract

Orientation estimation from magnetic, angular rate, and gravity (MARG) sensor array is a key problem in mechatronic-related applications. This paper proposes a new method in which a quaternion-based Kalman filter scheme is designed. The quaternion kinematic equation is employed as the process model. With our previous contributions, we establish the measurement model of attitude quaternion from accelerometer and magnetometer, which is later proved to be the fastest (computationally) one among representative attitude determination algorithms of such sensor combination. Variance analysis is later given enabling the optimal updating of the proposed filter. The algorithm is implemented on real-world hardware where experiments are carried out to reveal the advantages of the proposed method with respect to conventional ones. The proposed approach is also validated on an unmanned aerial vehicle during a real flight. Results show that the proposed one is faster than any other Kalman-based ones and even faster than some complementary ones while the attitude estimation accuracy is maintained.

Analysis of the Accuracy of Ten Algorithms for Orientation Estimation Using Inertial and Magnetic Sensing under Optimal Conditions: One Size Does Not Fit All

“…Moreover, the heading resulting from the magnetometer measurements needs to be dealt with care due to the presence of external magnetic fields. Researchers from different fields such as navigation and biomechanics have proposed several sensor fusion implementations over the years, including machine and deep learning approaches, to provide accurate orientation estimates using MIMUs [ 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 ]. The large majority of the published sensor fusion algorithms (SFAs) can be grouped in two main classes: Kalman filters (KF) [ 24 ] and complementary filters (CF).…”

Fast AHRS Filter for Accelerometer, Magnetometer, and Gyroscope Combination with Separated Sensor Corrections

“…To minimize the potential of introducing an error in implementation, we preferred to compare the quality of our filter with the methods published with implemented code. Therefore, we chose the Madgwick’s complementary filter [ 21 ], the Valenti’s complementary filter [ 18 ], and the Guo’s Fast Kalman Filter [ 16 ]. However, none of the recently published filters with independent corrections from the magnetometer and accelerometer had available code.…”

“…The performance benefit of these methods usually comes with a significant computational price, making it unsuitable for low cost, low-power applications. While multiple techniques decreasing computational cost of the Kalman filters exists [ 16 ], adaptation of covariance matrices is always expensive.…”

Design And Implementation Of A Real Time Attitude Estimation System With Low Cost Sensors

“…It is well known that through integration it is possible to obtain the attitude using the gyroscope measurements exclusively, but as will be see in the results, the angular-rate integral can diverge very quickly. In the literature some practical solutions include sensor fusion, for example, the complementary filters with constant gain presented in (Madgwick et al, 2011) and the adaptive gain in (Valenti et al, 2015), also the well known linear Kalman filter (KF) (Valenti et al, 2016;Guo et al, 2017;Feng et al, 2017) and the extended Kalman filter (EKF) (Sabatini, 2006;Baroni, 2017). Complementary Filters can be computationally more efficient, however, they rely on experimental results for tuning the gains and it can work well only in specific cases.…”