A Human Motion Tracking Algorithm Using Adaptive EKF Based on Markov Chain

In this paper, a precise attitude estimation of human body segments using a Markov chain-based adaptive Kalman filter is proposed. For the attitude estimation with inertial sensors and magnetometers, which is called to attitude and heading reference system (AHRS), a measurement disturbance mitigation algorithm is the most important part. In this paper, new parameters for detecting disturbances on accelerometers and magnetometers are applied to prevent divergence. In addition, for the rapid detection of disturbances, a hidden Markov-based adaptation rule is developed. By the proposed adaptive filter with these two major improvements, the disturbed measurement is effectively detected and mitigated, and experimental results show the improved attitude estimation performances over other AHRS algorithms.