Geomagnetic fingerprinting is a promising technology for supplying smartphone indoor navigation algorithms with infrastructureless and to some extent stable local position information. Geomagnetic disturbances in buildings impose a characteristic magnetic signature which can be detected by the phones magnetic sensor. A common approach is to create a database by recording reference fingerprints for example along a predefined known path. In the positioning phase, magnetic data is recorded along a certain time- or path length of the unknown path. The live data can be analyzed and compared against the prerecorded reference fingerprints. This magnetic matching procedure difers considerably from WiFi fingerprinting, where WiFi data from discrete points is compared. The main diference is that the described approach uses data recorded as time series. The matching has to consider not only the signal amplitude but also temporal- or spatial matching. In this paper several magnetic matching algorithms are evaluated for usage in an indoor positioning system. A public database is used as data source allowing comparison of the results with other works.
data, as well as barometric- and magnetic data. The obutained accuracy may then be increased
for example by Kalman filtering or application of a particle filter [
The paper focuses on using the geomagnetic field for indoor positioning [
In order to have a reliable data source a publicly available database was used for the analysis.
The database of the smartphone ofline Track from the IPIN 2018 competition has been chosen. It
contains more than 40 training- and validation data sets each covering 5..10 minutes of walking
annotated with true ground information of 41 diferent ground truth points, distributed on 3
lfoors of the Atlantis shopping mall, Nantes, France. Also a large 20 minute evaluation data set
with about 100 true ground references is available [
If the magnetic field is measured along a defined path in a building, it generates a characteristic
measurement curve. One possible principle of position determination based on the magnetic
ifeld works as follows: First, certain paths have to be defined in a building and the magnetic
ifeld has to be recorded along these paths and stored in a database. These recordings are the
magnetic fingerprints. If an agent now wants to determine his position, he also measures the
magnetic field along his path. For determination of the current position, the values for the last
meters of walking, for example the last 5 meters are used. Now he searches the database for
sections in the records (the fingerprints) that match his own recording. If such a section is
found, it is likely that the agent is at the position that belongs to the end of the found section.
The procedure is referred to as geomagnetic fingerprinting [
This method is very diferent from WiFi fingerprinting methods [
A matching procedure for magnetic fingerprints shall deliver a likelihood measure, e.g. a numerical value, and a displacement value e.g. a number expressing the temporal or spatial displacement between the magnetic trace from the evaluation path and a prerecorded reference path.
In the following sections, referring to the magnetic field or the magnetic signal means referring to the amplitude of the magnetic field vector. Of course, also other information like the z-component of the field or combinations could be employed for fingerprinting. There are two kinds of records used: In the positioning phase, the magnetic amplitude is recorded by the agents smartphone. This record, for which the true ground positions are to be determined, is named “evaluation trace”. The unknown path itself is referred to as the “evaluation path”. A magnetic reference measurement along a known path, recorded during the initial mapping phase, is referred to as reference fingerprint or “reference trace” (alias “training trace”). 3. Time to path mapping The database comprises numerous “long reference records” which, each in diferent well-defined paths, pass a subset of 41 known true ground reference points, and the passage of the points is marked in the record. By employing a PDR positioning algorithm [19, 20] continuous position estimates are performed for these reference records. Magnetic data is thereby mapped from elapsed time to passed path length as well as ground position. For each 10 cm of passed path length, a data point is generated. From each “long record”, new records are split of each time a known true ground reference position is passed. These reference records each are cut such that they start 40 meters before passing the reference point and end 10 meters after the reference point. That is also the x scale on the figures presented later.
The database comprises also a “long evaluation trace”. Also for this evaluation trace the time when the agent passes the reference points is known. For this paper, a representative cut of the trace has been selected to serve as base for the analysis. The cut comprises a 20 meter sequence ending before the selected reference point. In a practical scenario, the true ground and passed positions may not be known for an evaluation path and a relative mapping from time to passed path length is not generally possible. Therefore in the paper, the mapping is performed by using only the step counter and a fixed step length. This compensates for stops in the walk, but not for diferent walking speeds. 4. Evaluating of methods for comparing magnetic traces Figure 1 (a) shows 3 magnetic reference traces and one evaluation trace. All traces are high-pass ifltered to remove the constant field contribution and low-pass filtered to reduce signal noise. The x-axis shows the route distance to a certain reference point of the database. Only those traces aligned to the walking direction are shown. The evaluation trace has been also adjusted ) n iitoa 1 v e edd 0.5 u it lapm 0 d litefi -0.5 c aeng -1 m ( ]T -1.5 o rc i[m -2 B -40 80 (c) i)ts 60 n .rub 40 a ( itry 20 a l iim 0 S -20 (a)
Training and evaluation data for ref. 23 1. T05-01.txt 23(15) t= 274.8 2. T05-02.txt 23(15) t= 269.8 3. T05-03.txt 23(15) t= 272.5 4. EVALUATION.txt t= 712.9 7 (b)
Euclidian distance (train - eval) over offset sqsum 1 sqsum 2 sqsum 3 -30
-20 -10 s [m] (route distance to reference point)
Convolution folded 1 folded 2 folded 3 0 10 -2 -1.5 -1 s [m-]0(.c5orrelat0ion offs0e.t5) 1 1.5 2
Convolution Zoom folded 1 folded 2 folded 3 6 as the passing time of the reference point was known. In general, the path distance between reference and evaluation is not known but is the number to be determined. If, for example, the path distance between evaluation and a certain reference was -3 meters (3 meters before passing the reference point), the current position of the agent would be the position which is annotated for -3 m in the reference track. So, by determination of the route distance between reference and evaluation path, the unknown position can be determined.
Deviation from true position Euclidian dist. min value Euclidian dist. min. path [m] Convolution max. val.
Convolution max. path [m] DTW min value DTW min path [m] Modif. DTW min value Modif DTW min path [m]
The distance will be determined by correlation. In 1 (b) the Euclidean distance between a reference fingerprint and the evaluation fingerprint is shown over the fingerprint correlation original DTW fit
modified DTW fit warped 1 warped 2 warped 3 eval. Trace (d) warped 1 warped 2 warped 3 eval. Trace -20 -15 -10 -5 s s [m] (route distance to reference point) aaaaaaaaaass DTW mapping DTW mapping
RRReeefff 231 1 (c) RRReeefff 312 0 -20
-15 -10 -5 s [m] (route distance to reference point) DTW mapping DTW mapping
RRReeefff 231 1 (f) RRReeefff 231 0 (a) ) n itoa 0. 5 i v e eudd 0 lit p am-0.5 d l e fi itce -1 n g a (m-1.5 ] T o ircm -2 [ B
5 (b) ) 0 s k trca -5 g inn-10 ir a ton-15 e can-20 ts i(d-25 ] m [s-30 -35 0 -1 -2 -3 ) n itoa 0. 5 i v e eudd 0 ilt p am-0.5 d l e fi itce -1 n g a (m-1.5 ] T o ircm -2 [ B
5 (e) ) 0 s k trca -5 g inn-10 ir a ton-15 e can-20 ts i(d-25 ] m [s-30 -35 0 -1 -2 -3 ofset. The minimum x value indicates the path ofset in meters for the best match. In the ideal case this would be zero, as the measurements are aligned with the real time when each of the traces passes the reference point. Practically a small misalignment may be caused by measurement noise. Another source is the mapping of time to path length, which depends on the accuracy of the step detection. Nevertheless, the maximum obtained displacement is 0.5 m only, which is a reasonable accuracy.
Figure 1 (c) and (d) (zoomed) brings the result of a convolution. As expected, the results are quite comparable to the ones obtained by the Euclidean distance.
Figure 2 displays the result of dynamic time warping (DTW) (see for example [
Magnetic fingerprinting has been investigated on the IPIN2018 track 3 smartphone ofline database. By mapping magnetic readings to positions respectively passed path length, magnetic ifngerprints have been constructed. It has been shown that positions may be mapped to reference points with an accuracy of better than 0.5 meters, for an exemplary case (Table 1). Three diferent approaches have been compared. It was found that Euclidean distance and convolution as well as DTW deliver results with comparable accuracy. For DTW, a modification has been proposed that favors diagonal mapping by introduction of penalties. However, DTW seems to be not as stable as the others.
The next step will be to continue the integration of the magnetic fingerprint algorithm into an existing positioning algorithm. From the preliminary results we deduct that, at regions where feature-carrying magnetic data is available and is close to true-ground references, an unknown path may be matched with a spatial accuracy in the range of 1 meter or better. The positioning phase will be fully implemented to demonstrate and evaluate its feasibility in a lab setup and with publicly available evaluation data.
This work was funded by the Carl-Zeiss-Stiftung Stuttgart in project “SensAR” in the frame of the programme “transfer”. Conference on Indoor Positioning and Indoor Navigation, 2013, pp. 1–10. doi:10.1109/ IPIN.2013.6817853. [17] P. Robertson, M. Frassl, M. Angermann, M. Doniec, B. J. Julian, M. Garcia Puyol, M. Khider, M. Lichtenstern, L. Bruno, Simultaneous localization and mapping for pedestrians using distortions of the local magnetic field intensity in large indoor environments, in: International Conference on Indoor Positioning and Indoor Navigation, 2013, pp. 1–10. doi:10.1109/IPIN.2013.6817910. [18] S. Knauth, 8 - study and evaluation of selected rssi-based positioning algorithms, in: J. Conesa, A. Pérez-Navarro, J. Torres-Sospedra, R. Montoliu (Eds.), Geographical and Fingerprinting Data to Create Systems for Indoor Positioning and Indoor/Outdoor Navigation, Intelligent Data-Centric Systems, Academic Press, 2019, pp. 147 – 167. URL: http://www.sciencedirect.com/science/article/pii/B978012813189300006X. doi:https:// doi.org/10.1016/B978-0-12-813189-3.00006-X. [19] S. Knauth, A. Koukofikis, Smartphone positioning in large environments by sensor data fusion, particle filter and fcwc, in: 2016 International Conference on Indoor Positioning and Indoor Navigation (IPIN), 2016, pp. 1–5. doi:10.1109/IPIN.2016.7743706. [20] S. Knauth, Smartphone pdr positioning in large environments employing wifi, particle iflter, and backward optimization, in: 2017 International Conference on Indoor Positioning and Indoor Navigation (IPIN), 2017, pp. 1–6. doi:10.1109/IPIN.2017.8115918. [21] E. Keogh, C. A. Ratanamahatana, Exact indexing of dynamic time warping, Knowledge and Information Systems 7 (2005) 358–386. URL: https://doi.org/10.1007/s10115-004-0154-9. doi:10.1007/s10115-004-0154-9.