05 / 2023 – 09 / 2023
Localization and navigation outdoors is most commonly done by using global navigation satellite systems (GNSS). When it comes to pedestrian tracking indoors, GNSS can’t provide reliable and precise coverage due to its lack of accuracy inside buildings . Instead using pedestrian dead reckoning (PDR) for indoor navigation has been a commonly used approach, especially since nowadays inertial measurement units (IMU) have been widely integrated in most smart wearable devices . The classic PDR approach is to integrate gyroscope and accelerometer sensor data over time for each step taken to track the user’s heading and position. There have been many different studies using different sensors and body placements, as well as variations of the PDR algorithm itself. When it comes to the design process of PDR systems, several sub-problems need to be considered. These include issues such as sensor placement, different and complex human motion types, and the core problem of estimating the user’s trajectory . Most studies show high levels of tracking accuracy in test scenarios for the systems they have developed. However, testing conditions often do not reflect the complexity of human movement in use cases outside the lab. Therefore, it’s not certain that these models can achieve the same accuracy in real-world applications, and it’s very likely that errors will increase . A fundamental problem of integrating the sensor data, is the error accumulation over time. This means that even if the models have high tracking accuracy for short time periods, the absolute error will grow steadily . In an attempt to overcome this limitation, other PDR models have been developed that use deep learning to process raw IMU data to accomplish the task of regressing the full trajectory. This reduces the accumulation of errors over time, by avoiding the need to integrate sensor data and ideally the models will learn to deal with the sensor drift [5, 6].
 S. Beauregard and H. Haas, “Pedestrian dead reckoning: A basis for personal positioning,” Proceedings of the 3rd Workshop on Positioning, Navigation and Communication (WPNC’06), 01 2006.
 Y. Wu, H.-B. Zhu, Q.-X. Du, and S.-M. Tang, “A survey of the research status of pedestrian dead reckoning systems based on inertial sensors,” International Journal of Automation and Computing, vol. 16, pp. 1–19, 10 2018.
 X. Hou and J. Bergmann, “Pedestrian dead reckoning with wearable sensors: A systematic review,” IEEE Sensors Journal, vol. 21, no. 1, pp. 143–152, 2021
 M. Kok, J. D. Hol, and T. B. Sch¨on, “Using inertial sensors for position and orientation estimation,” Foundations and Trends® in Signal Processing, vol. 11, no. 1-2, pp. 1–153, 2017. [Online]. Available: https://doi.org/10.1561/2000000094
 S. Herath, H. Yan, and Y. Furukawa, “Ronin: Robust neural inertial navigation in the wild: Benchmark, evaluations, new methods,” in IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2020, pp. 3146–3152.
 O. Asraf, F. Shama, and I. Klein, “Pdrnet: A deep-learning pedestrian dead reckoning framework,” IEEE Sensors Journal, vol. 22, no. 6, pp. 4932–4939, 2022
 M. Edel and E. K¨oppe, “An advanced method for pedestrian dead reckoning using blstm-rnns,” in 2015 International Conference on Indoor Positioning and Indoor Navigation (IPIN), 2015, pp. 1–6.
 L. Xu, Z. Xiong, J. Liu, Z. Wang, and Y. Ding, “A novel pedestrian dead reckoning algorithm for multi-mode recognition based on smartphones,” Remote Sensing, vol. 11, p. 294, 02 2019.
 F. Ott, D. R¨ugamer, L. Heublein, B. Bischl, and C. Mutschler, “Joint classification and trajectory regression of online handwriting using a multi-task learning approach,” in 2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2022, pp. 1244–1254.
 “Test and application center l.i.n.k.” Nov 2021. [Online]. Available: https://www.iis.fraunhofer.de/en/profil/standorte/linkhalle.html