Abstract | ||
---|---|---|
Recently various indoor positioning techniques have been developed based on smartphone. However, most of them need external signals. In this paper a self-contained approach relying on built-in inertial sensors is implemented. Taking advantage of Pedestrian Dead Reckoning, it updates the current position by measuring the length and the heading of each step. Foremost the whole walking process is divided into segments, in which only straight walking is involved. After that the Feature Vectors are extracted for step detection. Specially, to cope with the instabilities caused by gait change, an equivalent Model Wave is created to substitute the original data. Finally, Particle Filter is employed for map matching. According to a group of experiments, our approach is as accurate as traditional positioning technique but shows more robustness. |
Year | DOI | Venue |
---|---|---|
2015 | 10.1109/WCNC.2015.7127812 | WCNC |
Keywords | Field | DocType |
feature vector,model wave simulating,moving variance analysis,particle filter,pedestrian dead reckoning | Computer vision,Feature vector,Computer science,Particle filter,Real-time computing,Robustness (computer science),Dead reckoning,Step detection,Inertial measurement unit,Artificial intelligence,Map matching,Indoor positioning system | Conference |
ISSN | Citations | PageRank |
1525-3511 | 1 | 0.38 |
References | Authors | |
14 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yi Sun | 1 | 23 | 2.47 |
Zhao, Yubin | 2 | 1 | 0.38 |
jochen schiller | 3 | 366 | 42.37 |