The Extended Kalman Filter With Reduced Computation Time for Pedestrian Dead Reckoning

The extended Kalman filter (EKF) has been used in resea-rch works on attitude estimation of the inertial measurement unit (IMU) and pedestrian dead reckoning (PDR). It is important to improve computation cost to estimate attitude of IMU and the accuracy of attitude estimation for the practical pedestrian navigation system with IMU. In this letter, the Kaisoku EKF (KEKF) is proposed to improve the computation time of the EKF while keeping the accuracy of estimating walking trajectory. The KEKF updates the state vector using the prediction equations, which can be calculated in low computation cost during mid-stance phase. In the experiments, it is verified that the KEKF can estimate attitude of IMU faster than the EKF while keeping the accuracy of estimating walking trajectory. The computation time of the KEKF in mid-stance phase can decrease about 14.031% comparing with the EKF. The averages of relative errors of estimated walking distance were 4.235% in the KEKF.