[Back]

M. Zhu, K. Zhang, F. Wu, W. Cartwright, G. Retscher:
"An Investigation of Pedestrian Positioning Algorithms Based on Integrated Lowcost INS; GPS and Magnetometer";
Talk: IGNSS 2007 Conference, Sydney, Australia; 2007-12-04 - 2007-12-06; in: "IGNSS 2007 Conference", (2007), Paper ID 93, 11 pages.

Despite of the advantages of integrated Inertial Navigation Systems (INS) and GPS, their application in pedestrian positioning is still a challenge. Instead, most of the pedestrian positioning systems use dead reckoning (DR) or integrated DR/GPS systems rather than integrated INS/GPS systems to avoid the significant biases and noises from the low-cost sensors. These DR systems, however, have a main drawback which is difficult to determine the distance travelled and the position from the pedestrian step detection when the pedestrian moves smoothly. It would be better to use low-cost INS and a new integration algorithm to estimate the position accurately during smooth movements of the pedestrian. The integration algorithm, however, must provide a robust estimation to tackle the noises and biases from the low-cost sensors.

In this paper a new algorithm is proposed for a low-cost integrated INS, GPS and magnetometer system. This algorithm uses the transformation matrix from the alignment stage and the magnetometers´ measurements determining the attitude in order to avoid the attitude estimation drifts generated by INS. It also uses the velocity measured by the GPS signals´ Doppler shifts measurements to correct the velocity determinations and uses the GPS position measurements to correct the drifts of the position estimations. A number of experiments were conducted in metropolitan Melbourne. The results show that an average positioning accuracy using this new algorithm of 4.3 m with a standard deviation 2.2 m can be achieved. The comparison of the results of different scenarios indicate that the constraints and corrections added into the new algorithm make the system more robust to the errors from the low-cost INS and give a higher position update rate than using GPS alone. These advantages make the integrated system ideal for pedestrian positioning applications.

Project Head Georg Gartner:
Ubiquitäre Kartographie für Fußgängernavigation