IMU-Based but Magnetometer-Free Joint Angle Estimation of Constrained Links

Abstract

Recently, there is an increasing interest in joint angle estimation methods using inertial measurement units (IMUs). To obtain the joint angle, the orientations of two adjacent links are required by using IMUs as well as magnetometers. However, when a magnetometer is used for heading estimation, magnetometer signals can seriously degrade accuracy of the orientation estimation in magnetically disturbed environments. This research proposes a novel IMU-based but magnetometer-free joint angle estimation method of constrained links. The key underlying concept of this research is, whereas orientation estimation for an unconstrained link unavoidably requires a magnetometer for heading determination, joint angle estimation of constrained links does not have to. A kinematic constraint may replace the information from the magnetometer. Hence, this paper presents a constraint-augmented Kalman filter based on IMU signals that eliminates magnetometer-related uncertainty for robust IMU-based joint angle estimation. Our experimental results show that the proposed method is superior to two other methods (l.e., 4.12° and 2.68° of estimation improvement from the two other methods).