Quaternion Vector Based Pedestrian Dead-Reckoning with Smartphones

Essiz G., SOYSAL G.

27th Signal Processing and Communications Applications Conference (SIU), Sivas, Türkiye, 24 - 26 Nisan 2019 identifier identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1109/siu.2019.8806544
  • Basıldığı Şehir: Sivas
  • Basıldığı Ülke: Türkiye
  • Anahtar Kelimeler: attitude estimation, quaternion, pedestrian dead-reckoning, step detection, adaptive step length estimation
  • Ankara Üniversitesi Adresli: Evet

Özet

Global Positioning System (GPS) is frequently used for location determination in military and civil aviation sector, vehicles, smartphones etc. However, when satellite signal is weak or there is not any signal to connect and also for indoor navigation problems, an alternative solution is needed for location estimation. This study contains Pedestrian Dead-Reckoning (PDR) by aid of smartphone body-mounted Micro Electro-Mechanic (MEMS) without GPS usage. Quaternion based attitude estimation is used to avoid Euler integration singularity problem. Furthermore, attitude estimation with accelerometer/magnetometer and the quaternion attitude outputs is fused with Kalman Filter (KF) to get rid of gyroscope drift error. Steps are detected by using accelerometer magnitude and then step frequencies are calculated for the adaptive step length estimation. Thereafter, pedestrian position is obtained by using the estimated attitude and step length. Indoor and outdoor test scenarios are specified and with gathering data, PDR is performed to designate the accuracy of this study.