Regresar

Orientation Estimation Using Filter-Based Inertial Data Fusion for Posture Recognition

Abstract:

In this article, the Kalman filter, Mahony filter and Madgwick filter are implemented to estimate the orientation from inertial data, using an IMU called 9 × 3 of the MoMoPa3 project which contain various sensors including a gyroscope, an accelerometer and a magnetometer, each one of them, equipped with three perpendicular axes, in the magnetometer the measurement was modified to correct the distortions by hard metals, demonstrating improvements in the accuracy of the orientation estimates. In addition, the Kinovea video analyzer software is used as reference and gold standard to calculate the Root-Mean-Square Error (RMSE) with each filter. When comparing the angles estimated by the filters with those obtained from Kinovea, it was observed that one of the filters was better in performance. The information obtained in this article can be involved in several fields of science, one of the most important in the field of medicine, helping to control Parkinson’s disease since it allows to evaluate and recognize when a patient suffers a fall or presents Freezing of the gait (FOG).