Page 45 - Fister jr., Iztok, and Andrej Brodnik (eds.). StuCoSReC. Proceedings of the 2018 5th Student Computer Science Research Conference. Koper: University of Primorska Press, 2018
P. 45
ure 2: Hand-reading position. same model by employing substitution [5]. The model pro-
posed by Do et al. [3] expectedly performed the worst since
performance as: it does not contain any tuneable constants.
e = |d − d∗| · 100% (3) 7. CONCLUSION
d∗
In this work we presented our research towards an enhanced
where e is error, d is the estimated distance of walked path inertial sensor-based step length estimation model. We based
and d∗ the real distance of walked path. our model on the model proposed by Tian et al. [2] since we
obtained the best evaluation results for it for universal sets
5. RESULTS of constants [5].
Table 1 shows the performance of the models. It includes We eliminated user’s height from the model and included
mean errors and standard deviations of the models with mean absolute acceleration in walking direction instead. We
respect to total walked distance. Mean errors range from tested the proposed model for one sensor position in hand
3.76 % to 16.84 % and standard deviations from 2.41 % to and obtained promising evaluation results comparable to re-
7.56 %. lated models.
Table 1: The performance of the models. We will further improve the accuracy of the proposed model
and test it for more sensor positions. We also plan to modify
Models Mean Standard the proposed model, so it would require less time for tun-
errors [%] deviations [%] ing and thoroughly address the pre-processing of the sensor
data. Moreover, we will conduct additional experiments us-
Tian et al. [2] 4.35 2.78 ing different IoT devices and test the proposed model for
pedestrian-dead-reckoning-based indoor positioning.
Zhang et al. [1] 8.91 6.86
Acknowledgments
Renaudin et al. [6] 8.91 6.86
M. V. received funding for doctoral studies from the Univer-
Do et al. [3] 16.84 7.56 sity of Ljubljana – 2016 generation. Authors would like to
thank all the volunteers who participated in the experiment.
Weinberg [7] 4.36 3.80
8. REFERENCES
Diaz and 5.25 4.54
Gonzales [8] [1] P. Zhang, X. Chen, X. Ma, Y. Wu, H. Jiang, D. Fang,
Z. Tang, and Y. Ma, “SmartMTra: Robust Indoor
Mikov et al. [4] 7.44 5.93 Trajectory Tracing Using Smartphones,” IEEE Sensors
Journal, vol. 17, no. 12, pp. 3613–3624, 2017.
Bylemans et al. [9] 5.68 4.61
[2] Q. Tian, Z. Salcic, K. I. K. Wang, and Y. Pan, “A
Proposed model 3.76 2.41 Multi-Mode Dead Reckoning System for Pedestrian
Tracking Using Smartphones,” IEEE Sensors Journal,
6. DISCUSSION vol. 16, no. 7, pp. 2079–2093, 2016.
Overall, the proposed model slightly outperformed all the [3] T. N. Do, R. Liu, C. Yuen, M. Zhang, and U. X. Tan,
other models. Quite similar performance achieved the mod- “Personal Dead Reckoning Using IMU Mounted on
els proposed by the Tian et al. [2], Weinberg [7], Diaz and Upper Torso and Inverted Pendulum Model,” IEEE
Gonzales [8], Bylemans et al. [9] and Mikov et al. [4]. Even Sensors Journal, vol. 16, no. 21, pp. 7600–7608, 2016.
though the model proposed by Mikov et al. [4] extends the
Weinberg model [7] it did not perform better on average. [4] A. Mikov, A. Moschevikin, A. Fedorov, and A. Sikora,
The models proposed by Zhang et al. [1] and Renaudin et “A localization system using inertial measurement units
al. [6] performed similarly since they can be rewritten to the from wireless commercial hand-held devices,” in 2013
International Conference on Indoor Positioning and
Indoor Navigation (IPIN), 2013, pp. 1–7.
[5] M. Vezoˇcnik and M. B. Juriˇc, “Inertial Sensor-Based
Step Length Estimation Models: A Review,” In Review
Process, 2018.
[6] V. Renaudin, M. Susi, and G. Lachapelle, “Step Length
Estimation Using Handheld Inertial Sensors,” Sensors,
vol. 12, no. 7, pp. 8507–8525, 2012.
[7] H. Weinberg, “Using the ADXL202 in pedometer and
personal navigation applications,” Analog Devices
AN-602 application note, vol. 2, no. 2, pp. 1–6, 2002.
[8] E. M. Diaz and A. L. M. Gonzalez, “Step detector and
step length estimator for an inertial pocket navigation
system,” in 2014 International Conference on Indoor
Positioning and Indoor Navigation (IPIN), 2014, pp.
105–110.
StuCoSReC Proceedings of the 2018 5th Student Computer Science Research Conference 45
Ljubljana, Slovenia, 9 October
posed by Do et al. [3] expectedly performed the worst since
performance as: it does not contain any tuneable constants.
e = |d − d∗| · 100% (3) 7. CONCLUSION
d∗
In this work we presented our research towards an enhanced
where e is error, d is the estimated distance of walked path inertial sensor-based step length estimation model. We based
and d∗ the real distance of walked path. our model on the model proposed by Tian et al. [2] since we
obtained the best evaluation results for it for universal sets
5. RESULTS of constants [5].
Table 1 shows the performance of the models. It includes We eliminated user’s height from the model and included
mean errors and standard deviations of the models with mean absolute acceleration in walking direction instead. We
respect to total walked distance. Mean errors range from tested the proposed model for one sensor position in hand
3.76 % to 16.84 % and standard deviations from 2.41 % to and obtained promising evaluation results comparable to re-
7.56 %. lated models.
Table 1: The performance of the models. We will further improve the accuracy of the proposed model
and test it for more sensor positions. We also plan to modify
Models Mean Standard the proposed model, so it would require less time for tun-
errors [%] deviations [%] ing and thoroughly address the pre-processing of the sensor
data. Moreover, we will conduct additional experiments us-
Tian et al. [2] 4.35 2.78 ing different IoT devices and test the proposed model for
pedestrian-dead-reckoning-based indoor positioning.
Zhang et al. [1] 8.91 6.86
Acknowledgments
Renaudin et al. [6] 8.91 6.86
M. V. received funding for doctoral studies from the Univer-
Do et al. [3] 16.84 7.56 sity of Ljubljana – 2016 generation. Authors would like to
thank all the volunteers who participated in the experiment.
Weinberg [7] 4.36 3.80
8. REFERENCES
Diaz and 5.25 4.54
Gonzales [8] [1] P. Zhang, X. Chen, X. Ma, Y. Wu, H. Jiang, D. Fang,
Z. Tang, and Y. Ma, “SmartMTra: Robust Indoor
Mikov et al. [4] 7.44 5.93 Trajectory Tracing Using Smartphones,” IEEE Sensors
Journal, vol. 17, no. 12, pp. 3613–3624, 2017.
Bylemans et al. [9] 5.68 4.61
[2] Q. Tian, Z. Salcic, K. I. K. Wang, and Y. Pan, “A
Proposed model 3.76 2.41 Multi-Mode Dead Reckoning System for Pedestrian
Tracking Using Smartphones,” IEEE Sensors Journal,
6. DISCUSSION vol. 16, no. 7, pp. 2079–2093, 2016.
Overall, the proposed model slightly outperformed all the [3] T. N. Do, R. Liu, C. Yuen, M. Zhang, and U. X. Tan,
other models. Quite similar performance achieved the mod- “Personal Dead Reckoning Using IMU Mounted on
els proposed by the Tian et al. [2], Weinberg [7], Diaz and Upper Torso and Inverted Pendulum Model,” IEEE
Gonzales [8], Bylemans et al. [9] and Mikov et al. [4]. Even Sensors Journal, vol. 16, no. 21, pp. 7600–7608, 2016.
though the model proposed by Mikov et al. [4] extends the
Weinberg model [7] it did not perform better on average. [4] A. Mikov, A. Moschevikin, A. Fedorov, and A. Sikora,
The models proposed by Zhang et al. [1] and Renaudin et “A localization system using inertial measurement units
al. [6] performed similarly since they can be rewritten to the from wireless commercial hand-held devices,” in 2013
International Conference on Indoor Positioning and
Indoor Navigation (IPIN), 2013, pp. 1–7.
[5] M. Vezoˇcnik and M. B. Juriˇc, “Inertial Sensor-Based
Step Length Estimation Models: A Review,” In Review
Process, 2018.
[6] V. Renaudin, M. Susi, and G. Lachapelle, “Step Length
Estimation Using Handheld Inertial Sensors,” Sensors,
vol. 12, no. 7, pp. 8507–8525, 2012.
[7] H. Weinberg, “Using the ADXL202 in pedometer and
personal navigation applications,” Analog Devices
AN-602 application note, vol. 2, no. 2, pp. 1–6, 2002.
[8] E. M. Diaz and A. L. M. Gonzalez, “Step detector and
step length estimator for an inertial pocket navigation
system,” in 2014 International Conference on Indoor
Positioning and Indoor Navigation (IPIN), 2014, pp.
105–110.
StuCoSReC Proceedings of the 2018 5th Student Computer Science Research Conference 45
Ljubljana, Slovenia, 9 October
