[time]
2020-08-26T16:47:34+02:00

[team_name]
Indora

[team_institution]
Institute of Computer Science, P.J. Šafárik University, Faculty of Science, Jesenná 5, 041 54 Košice, Slovakia

[logolink]
n/a

[system_name]
indora

[website]
indora.ics.upjs.sk

[track]
3

[reference_person]
Miroslav Opiela

[email]
miroslav.opiela@upjs.sk

[description]
Short version (copy from last year):

Project INDORA originated from a few student research projects and final theses, especially the author's dissertation. Various bachelor and master theses have reviewed different aspects of a comprehensive indoor positioning system during recent years. Pedestrian dead-reckoning, map model and bayesian filtering are essential components of the proposed localization system. The main research focus is on a low-dimensional grid-based bayesian filtering, as a less elaborated alternative to Kalman and Particle filters widely used for the positioning. A semi-automatically generated map model helps to reduce the localization error introduced by noisy sensor measurements and inaccurate system configuration, e.g., a step length estimation. The considered use case involves the user with handheld smartphone. It relies mostly on the map and no additional infrastructure is required in the building. The floor transition detection based on barometer measurements allows the system to estimate the user position within a single floor.

Longer version (tailored for this competition):
The localization system developed by INDORA team from P.J.Šafárik University from Slovakia is focused on the use case covered by the competition Track 3. Core components for the positioning are PDR, map model, and bayesian filtering. The Bayesian filtering research is in the center of attention. Grid-based implementations are the topic of the competitor's dissertation (Miroslav Opiela). One of these filters was applied on IPIN 2018 on-site and IPIN 2019 both on-site and off-site tracks. Meanwhile, a few bachelor and master theses at our institute elaborate different aspects of the comprehensive indoor positining including floor transition, human activity recognition, navigation paths, map models, etc. The bachelor thesis of Jakub Džama was devoted to exploration of features, advantages and drawbacks of the particle filter. He is currently working on the master thesis with the title Smartphone-based indoor navigation application. The evaluation of the localization module would be performed on IPIN 2020 Track 3. The final method will be the elaboration of previous work. Existing support tools will be used to compare different approaches and to choose the most suitable version based on the results from the training and validation datasets and the visualization and the analysis of the evaluation logfile path. The key scheme will be the same as for previous competitions but all components will be revised for the improvement. The floor transition is detected using barometer measurements. The grid-based Bayes filter or the particle filter handle the inaccuracy introduced by noisy sensors and imperfect outputs from other components. Experiments are performed with the step length estimation. The fixed value is replaced by a calculated one (based on current research results in this domain) or incorporated in the particle filter (based on the aforementioned bachelor thesis). The map model bounds the possible locations and improves the accuracy. The same approach is used as for previous competitions. We plan to investigate how to incorporate other sensors to the solution. Experiments with the sensor fusion using the light sensor improves the ability to distinguish between two similar corridors last year (corridor vs. balcony) and detect indoor-outdoor transitions. We elaborated the process of integration of other inputs to grid-based filters or to the selected versions of the particle filter. We plan to analyse the possibility to utilize additional sources of inputs, e.g., BLE, Wi-Fi, or light sensor. 

[references]
* Galčík, F., 

[Opiela,_M__(2016,_October)__Grid-based_indoor_localization_using_smartphones__In_2016_International_Conference_on_Indoor_Positioning_and_Indoor_Navigation_(IPIN)_(pp__1-8)__IEEE_

*_Renaudin,_V_,_Ortiz,_M_,_Perul,_J_,_Torres-Sospedra,_J_,_Jiménez,_A__R_,_Pérez-Navarro,_A_,_____]


[Ben-Moshe,_B__(2019)__Evaluating_indoor_positioning_systems_in_a_shopping_mall:_The_lessons_learned_from_the_IPIN_2018_competition__IEEE_Access,_7,_148594-148628_

*_Jakub_Džama__Indoor_localization_using_particle_filter,_2019__Bachelor’s_Thesis_

*_Galčík,_F_,_]


[Opiela,_M__Grid-Based_Bayesian_Filtering_Methods_for_Pedestrian_Dead_Reckoning_Indoor_Positioning_Using_Smartphones_(under_review_in_MDPI_Sensors_Journal)]