Drift-Free and Self-Aligned IMU-based Human Gait Tracking System with Augmented Precision and Robustness

摘要

IMU-based human joint motion acquisition system is attractive for real-time control and monitoring in the emerging wearable technology due to its portability. However, in practical applications, it heavily suffers from long-term drift, magnetic interference and inconsistency of rotational reference frames, which causes precision degradation. In this letter, a novel on-line IMU-based human gait estimation framework was proposed to obtain the joint rotational angles directly under the kinematic constraints between multiple body segments, whereas traditional methods need to estimate the orientation of each individual segment. This framework consists of an on-line algorithm to align IMU frames with human joints and motion estimation algorithms for hip and knee without the aid of magnetometer. Both a 2-DoF robot and human gait tests were performed to validate the proposed method as compared with the predictions from commercial IMUs, joint encoders and an optical tracking system. The outcome demonstrated its advantages of adaptive alignment, drift rejection and low computational cost, which alleviates the practical barriers faced by human motion data collection in the wearable devices.