Study on underwater navigation system using discrete hartley transform unscented Kalman filter for attitude estimation

To reduce the attitude estimate error and improve the calculation efficiency for navigation system applied to underwater glider, this paper proposes the discrete Hartley transform unscented Kalman filter (DUKF) utilizing inertial measuring units (IMUs). DUKF method uses efficient numerical calculation of the discrete Hartley transform and takes optimal estimate advantage of UKF in the nonlinear system to achieve fast and accurate attitude estimation. A real vehicle experiment has been done and the performance of proposed DUKF is compared with traditional EKF and UKF. The results show that the proposed DUKF outperforms EKF and UKF comprehensively taking account of estimation accuracy and computational burden.