Experimental Results

In our experimental results, we find out that in the detection of elevator, we are able to detect all the patterns we have recorded in training phase as Fig. 6.3. Also, in the detection of stairs, we can find out that three methods we propose have overall accuracy over 85% as Fig. 6.4 and Fig. 6.5. Also, we can also conclude from the chart that longest common subsequence has even better result over the other 2 methods. However, when looking at up detection accuracy, we can find out that pattern matching and longest common subsequence is fairly good, even better than longest common subsequence, but when looking at down accuracy, we can know that longest common subsequence is better than the other 2 methods since even though walking and walking downstairs is relatively similar, longest common subsequence can still have good determination.

Figure 6.3: Graph: Performance of Elevator Detection.

Figure 6.4: Graph: Performance of Stairs Detection.

Algorithm Pattern Matching LC Substring LC Subsequence

Figure 6.5: Chart: Performance of Stairs Detection.

Chapter 7 Conclusion

We have proposed a human behavior recognition system based on IMU sensors. Formerly, in order to solve the drifting problem in indoor localization, especially in multistory environment, altitude information is required. In order to obtain altitude information, several methods was proposed, such as using accelerometer or barometer. However, these methods are not precise and some may need extra hardware or infrastructure. Therefore, we proposed a solution to solve this problem by recognizing human behavior especially of changing floors such as taking elevator going up or down and walking upstairs or downstairs. Our system needs only IMU sensors to measures acceleration and Euler angles to detect human behavior of changing floor.

According to our observations, repeated patterns are found in walking upstairs and downstairs and similar pattern are found when taking elevator going up or down. This way, we record the meaningful patterns in advanced in our training process. Then, in online process, we are able to match the current pattern with training patterns. We proposed 3 methods of matching pat-terns including pattern matching, longest common substring and longest common subsequence.

There three methods are proposed to fit various pattern of human behavior and may match pattern in considering of time and magnitude constraint. Therefore in order to assist indoor localization system, we can lock the user in the same floor to avoid drifting between floors and change user’s floor when respecting behavior is triggered.

Bibliography

[1] P. Bahl and V. Padmanabhan. Radar: an in-building rf-based user location and tracking system.

volume 2, pages 775 –784 vol.2, 2000.

[2] S. Beauregard. Omnidirectional pedestrian navigation for first responders. pages 33 –36, mar.

2007.

[3] S. Beauregard and H. Haas. Pedestrian dead reckoning: A basis for personal positioning. In Proceedings of the 3rd Workshop on Positioning, Navigation and Communication(WPNC06), 2006.

[4] O. Bebek, M. A. Suster, S. Rajgopal, M. J. Fu, X. Huang, M. C. Cavusoglu, D. J. Young, M. Mehre-gany, A. J. van den Bogert, and C. H. Mastrangelo. Personal navigation via high-resolution gait-corrected inertial measurement units. Instrumentation and Measurement, IEEE Transactions on, PP(99):1 –10, 2010.

[5] I. Constandache, R. Choudhury, and I. Rhee. Towards mobile phone localization without war-driving. pages 1 –9, mar. 2010.

[6] L. Fang, P. Antsaklis, L. Montestruque, M. McMickell, M. Lemmon, Y. Sun, H. Fang, I. Koutroulis, M. Haenggi, M. Xie, and X. Xie. Design of a wireless assisted pedestrian dead reckoning system - the navmote experience. Instrumentation and Measurement, IEEE Transactions on, 54(6):2342 – 2358, dec. 2005.

[7] W. T. Faulkner, R. Alwood, D. W. A. Taylor, and J. Bohlin. Altitude accuracy while tracking pedestrians using a boot-mounted imu. pages 90 –96, may. 2010.

[8] E. G. G.-B. J. FELIZ ALONSO, Rau’l; ZALAMA CASANOVA. Pedestrian tracking using inertial sensors. Journal of Physical Agents, 3(1):35 – 43, jan. 2009.

[9] C. Fischer and H. Gellersen. Location and navigation support for emergency responders: A survey.

Pervasive Computing, IEEE, 9(1):38 –47, jan. 2010.

[10] E. Foxlin. Pedestrian tracking with shoe-mounted inertial sensors. Computer Graphics and Appli-cations, IEEE, 25(6):38 –46, nov. 2005.

[11] G. Glanzer, T. Bernoulli, T. Wiessflecker, and U. Walder. Semi-autonomous indoor positioning using mems-based inertial measurement units and building information. pages 135 –139, mar.

2009.

[12] A. Jimenez, F. Seco, C. Prieto, and J. Guevara. A comparison of pedestrian dead-reckoning algo-rithms using a low-cost mems imu. pages 37 –42, aug. 2009.

[13] O. M. Jussi Collin and G. Lachapelle. Indoor positioning system using accelerometry and high accuracy heading sensors. In Proceedings of GPS/GNSS 2003 Conference (Session C3), 2003.

[14] M. Kourogi, T. Ishikawa, and T. Kurata. A method of pedestrian dead reckoning using action recognition. pages 85 –89, may. 2010.

[15] J. McLellan, J. Schleppe, D. McLintock, and G. Deren. Gps/barometry height-aided positioning system. pages 369 –375, apr. 1994.

[16] MicroStrain. 3dm-gx1 datasheet. http://www.microstrain.com/pdf/

3DM-GX1DatasheetRev1.pdf.

[17] L. Ojeda and J. Borenstein. Personal dead-reckoning system for gps-denied environments. pages 1 –6, sep. 2007.

[18] J. Parviainen, J. Kantola, and J. Collin. Differential barometry in personal navigation. pages 148 –152, may. 2008.

[19] Y. S. Suh and S. Park. Pedestrian inertial navigation with gait phase detection assisted zero velocity updating. pages 336 –341, feb. 2009.

[20] R. L. R. Thomas H. Cormen, Charles E. Leiserson and C. Stein. 15.4, Introduction to Algorithms.

MIT Press and McGraw-Hill, Reading, Massachusetts, 2nd edition, 2001.

[21] H. Wang, H. Lenz, A. Szabo, J. Bamberger, and U. Hanebeck. Wlan-based pedestrian tracking using particle filters and low-cost mems sensors. pages 1 –7, mar. 2007.

[22] O. Woodman and R. Harle. Rf-based initialisation for inertial pedestrian tracking. In H. Tokuda, M. Beigl, A. Friday, A. Brush, and Y. Tobe, editors, Pervasive Computing, volume 5538 of Lecture Notes in Computer Science, pages 238–255. Springer Berlin / Heidelberg.

[23] O. Woodman and R. Harle. Pedestrian localisation for indoor environments. In UbiComp ’08:

Proceedings of the 10th international conference on Ubiquitous computing, pages 114–123, New York, NY, USA, 2008. ACM.