A Personal Navigation System Using Low-Cost MEMS/GPS/Fluxgate

The advances of MEMS technology have led the development of small-sized sensors (inertial sensors, magnetic compass sensors, etc.). Small-sized GPS receivers also have been developed using only one GPS chipset. For the synergistic effect, the integrated system of these systems has been adopted in many navigation systems. In this paper, a new algorithm for pedestrian navigation systems (PNS) using the integrated system is proposed. The key of the PNS is step (stance phase) detection. The step is detected using the accelerometer signals. The point of time for calculating of the inclination of the ground and the azimuth is determined through the step detection. The stride is computed using a neural network. The neural network is learned when the GPS signal is available. And then the walking distance and the position of the user are calculated. The performance of the proposed approach is verified through the actual walking test. With position compensation about every 100 steps, about 160m, the proposed system is able to keep an accuracy of 10m.