We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Unsupervised Movement Detection in Indoor Positioning Systems

Abstract: In recent years, the usage of indoor positioning systems for manufacturing processes became increasingly popular. Typically, the production hall is equipped with satellites which receive position data of sensors that can be pinned on components, load carriers or industrial trucks. This enables a company e.g. to reduce search efforts and to optimize individual system processes. In our research context, a sensor only sends position information when it is moved. However, various circumstances frequently affect that data is undesirably sent, e.g. due to disrupting factors nearby. This has a negative impact on the data quality, the energy consumption, and the reliability of the whole system. Motivated by this, we aim to distinguish between actual movements and signals that were undesirably sent which is in particular challenging due to the susceptibility of indoor systems in terms of noise and measuring errors. Therefore, we propose two novel unsupervised classification algorithms suitable for this task. Depending on the question of interest, they rely either on a distance-based or on a time-based criterion, which allows to make use of all essential information. Furthermore, we propose an approach to combine both classifications and to aggregate them on spatial production areas. This enables us to generate a comprehensive map of the underlying production hall with the sole usage of the position data. Aside from the analysis and detection of the underlying movement structure, the user benefits from a better understanding of own system processes and from the detection of problematic system areas which leads to a more efficient usage of positioning systems. Since all our approaches are constructed with unsupervised techniques, they are handily applicable in practice and do not require more information than the output data of the positioning system.
Subjects: Machine Learning (cs.LG); Applications (stat.AP)
Cite as: arXiv:2109.10757 [cs.LG]
  (or arXiv:2109.10757v1 [cs.LG] for this version)

Submission history

From: Jonathan Flossdorf [view email]
[v1] Sat, 21 Aug 2021 10:30:09 GMT (1341kb,D)

Link back to: arXiv, form interface, contact.