Gait-Based Authentication Using Anomaly Detection with Acceleration of Two Devices in Smart Lock

  • Kazuki Watanabe
  • Makoto Nagatomo
  • Kentaro Aburada
  • Naonobu Okazaki
  • Mirang ParkEmail author
Conference paper
Part of the Lecture Notes in Networks and Systems book series (LNNS, volume 97)


Currently, authentication in Smart locks is performed by fingerprint or face authentication. However, these authentications are inconvenient for smart locks because they require the user to stop for several seconds in front of the door and remove certain accessories (e.g., gloves, sunglasses). In this paper, we propose a user authentication method based on gait features. We propose a system model of gait-based authentication method using accelerometers in a smartphone and a wearable device (i.e., smartwatch), that is robust for unknown data using anomaly detection by machine learning. In addition, we conduct experiment to confirm the authentication rate of the proposed gait-based authentication. As a result, when using Isolation Forest as the anomaly detection algorithm, the average FAR (False Acceptance Rate) was 8.3%, the average FRR (False Rejection Rate) was 9.5%. Furthermore, we found that the better algorithm of anomaly detection of FAR and FRR is different depending on the subjects.



This work was supported by JSPS KAKENHI Grant Numbers JP17H01736, JP17K00139.


  1. 1.
    Ministry of Internal Affairs and Communications: 2018 White Paper on Information and Communication in Japan (2018).
  2. 2.
    Gartner: Gartner Says Worldwide Wearable Device Sales to Grow 26 Percent in 2019. Accessed 30 July 2019
  3. 3.
    Qrio: Qrio Smart Lock. Accessed 30 July 2019
  4. 4.
    August: August Smart Lock. Accessed 30 July 2019
  5. 5.
    Kwikset: Door Locks Door Hardware Smart Locks & Smart key Technology. Accessed 30 July 2019
  6. 6.
    Muaaz, M., Mayrhofer, R.: Smartphone-based gait recognition: from authentication to imitation. IEEE Trans. Mob. Comput. 16(11), 3209–3221 (2017)CrossRefGoogle Scholar
  7. 7.
    Hou, R., Watanabe, Y.: A Study on authentication at the time of the walk of using the acceleration sensor of smartphone. In: Computer Security Symposium, vol. 2013, pp. 21–23 (2013). (in Japanese)Google Scholar
  8. 8.
    Konno, S., Nakamura, Y., Shiraishi, Y., Takahashi, O.: Improvement of gait-based authentication by using multiple wearable sensors. IPSJ J. 57(1), 109–122 (2016). (in Japanese)Google Scholar
  9. 9.
    Iwamoto, T., Sugimori, D., Matsumoto, M.: A Study of identification of pedestrian by using 3-axis accelerometer. IPSJ J. 55(2), 734–749 (2014)Google Scholar
  10. 10.
    Mondal, S., Nandy, A., Chakraborty, P., et al.: Gait based personal identification system using rotation sensor. J. Emerg. Trends Comput. Inf. Sci. 3(3), 395–402 (2012)Google Scholar
  11. 11.
    Scikit-learn: scikit-learn machine learning in Python Scikit-learn 0.19.1 documentation. Accessed 30 July 2019

Copyright information

© Springer Nature Switzerland AG 2020

Authors and Affiliations

  • Kazuki Watanabe
    • 1
  • Makoto Nagatomo
    • 1
  • Kentaro Aburada
    • 2
  • Naonobu Okazaki
    • 2
  • Mirang Park
    • 1
    Email author
  1. 1.Kanagawa Institute of TechnologyAtsugiJapan
  2. 2.University of MiyazakiMiyazakiJapan

Personalised recommendations