The development of constraint algorithms and real-time smoothing for pedestrian indoor navigation with smartphones

Personal navigation technologies have been evolving rapidly because of the increasing popularity of smartphones and portable devices in recent years. This phenomenon has changed the developmental trend for pedestrian navigation with the use of available sensors. However, some considerations restrict the cost, accuracy and practicality of pedestrian navigation, especially in indoor environment without the Global Navigation Satellite System (GNSS). The Inertial Navigation System (INS) has been widely used for vehicle navigation. However, INS is inappropriate for pedestrian navigation because the complicated motions of human body causes the dynamic misalignment between the body frame and pedestrian frame. Thus, the human motion will produce the large navigation error after the integration of acceleration and angle rate during INS mechanization with vibration noise and misalignment error. Therefore, the INS mechanization requires the external measurements and constraints to improve the navigation accuracy, especially for real-time pedestrian indoor navigation. In general, the loosely couple is an ideal form for aiding the INS mechanization by integrate the position and orientation. This study uses the position of control point as the update measurement in the extended Kalman Filter (EKF) to aid the INS mechanization without GNSS. The sources of those control points can be pre-surveyed control points, map database, feature points and other Radio Frequency (RF) based indoor positioning technologies. This study proposes the Walking Velocity Constraint (WVC) to enable the use of walking velocity from a pedometer and step length model to constrain the velocity of moving direction in the EKF. Moreover, the real-time smoothing is also implemented with the use of EKF and control points. The performances of proposed algorithms are verified for pedestrian indoor navigation with smartphones. The preliminary results show the proposed algorithms improve the performance of INS in real-time for pedestrian navigation in indoor environment significantly.