The motion recognition based 3D indoor navigation system is developed to track the position of pedestrian in the indoor environment. The motion of pedestrian could be estimated by using the output of sensors embedded in the smartphone. To compensate the previous PDR system, we apply the estimated motion to PDR and make more accurate system than before. To enhance the availability of pattern recognition for motion and heading estimation, we develop the hybrid model and this enable the estimation of the motion and heading of various pedestrians. The position of pedestrian could be tracked by the proposed system though the pedestrian performs various motions.

Keywords: indoor navigation, pattern matching, pedestrian dead reckoning, smartphone, hybrid model