DC FieldValueLanguage
dc.contributor.authorVenzke, Marcus-
dc.contributor.authorKlisch, Daniel-
dc.contributor.authorKubik, Philipp-
dc.contributor.authorAli, Asad-
dc.contributor.authorDell Missier, Jesper-
dc.contributor.authorTurau, Volker-
dc.date.accessioned2020-12-18T11:36:45Z-
dc.date.available2020-12-18T11:36:45Z-
dc.date.issued2020-12-15-
dc.identifier.citationarXiv: 2012.08403v1 (2020)de_DE
dc.identifier.urihttp://hdl.handle.net/11420/8295-
dc.description.abstractIn this paper we investigate the usage of machine learning for interpreting measured sensor values in sensor modules. In particular we analyze the potential of artificial neural networks (ANNs) on low-cost micro-controllers with a few kilobytes of memory to semantically enrich data captured by sensors. The focus is on classifying temporal data series with a high level of reliability. Design and implementation of ANNs are analyzed considering Feed Forward Neural Networks (FFNNs) and Recurrent Neural Networks (RNNs). We validate the developed ANNs in a case study of optical hand gesture recognition on an 8-bit micro-controller. The best reliability was found for an FFNN with two layers and 1493 parameters requiring an execution time of 36 ms. We propose a workflow to develop ANNs for embedded devices.en
dc.language.isoende_DE
dc.subjectComputer Science - Learningde_DE
dc.subjectComputer Science - Learningde_DE
dc.subjectComputer Science - Neural and Evolutionary Computingde_DE
dc.subjectMachine Learningde_DE
dc.subjectArtificial Neural Networksde_DE
dc.subjectEmbedded Systemsde_DE
dc.subjectHand Gesture Recognitionde_DE
dc.subject.ddc000: Allgemeines, Wissenschaftde_DE
dc.titleArtificial Neural Networks for Sensor Data Classification on Small Embedded Systemsde_DE
dc.typePreprintde_DE
dc.type.dinipreprint-
dcterms.DCMITypeText-
tuhh.abstract.englishIn this paper we investigate the usage of machine learning for interpreting measured sensor values in sensor modules. In particular we analyze the potential of artificial neural networks (ANNs) on low-cost micro-controllers with a few kilobytes of memory to semantically enrich data captured by sensors. The focus is on classifying temporal data series with a high level of reliability. Design and implementation of ANNs are analyzed considering Feed Forward Neural Networks (FFNNs) and Recurrent Neural Networks (RNNs). We validate the developed ANNs in a case study of optical hand gesture recognition on an 8-bit micro-controller. The best reliability was found for an FFNN with two layers and 1493 parameters requiring an execution time of 36 ms. We propose a workflow to develop ANNs for embedded devices.de_DE
tuhh.publication.instituteTelematik E-17de_DE
tuhh.type.opusPreprint (Vorabdruck)-
dc.type.driverpreprint-
dc.type.casraiOther-
dc.identifier.arxiv2012.08403v1de_DE
local.status.inpressfalsede_DE
datacite.resourceTypeOther-
datacite.resourceTypeGeneralText-
item.creatorOrcidVenzke, Marcus-
item.creatorOrcidKlisch, Daniel-
item.creatorOrcidKubik, Philipp-
item.creatorOrcidAli, Asad-
item.creatorOrcidDell Missier, Jesper-
item.creatorOrcidTurau, Volker-
item.grantfulltextnone-
item.creatorGNDVenzke, Marcus-
item.creatorGNDKlisch, Daniel-
item.creatorGNDKubik, Philipp-
item.creatorGNDAli, Asad-
item.creatorGNDDell Missier, Jesper-
item.creatorGNDTurau, Volker-
item.mappedtypePreprint-
item.openairecristypehttp://purl.org/coar/resource_type/c_816b-
item.fulltextNo Fulltext-
item.cerifentitytypePublications-
item.languageiso639-1en-
item.openairetypePreprint-
crisitem.author.deptTelematik E-17-
crisitem.author.deptTelematik E-17-
crisitem.author.orcid0000-0002-3586-5878-
crisitem.author.orcid0000-0001-9964-8816-
crisitem.author.parentorgStudiendekanat Elektrotechnik, Informatik und Mathematik (E)-
crisitem.author.parentorgStudiendekanat Elektrotechnik, Informatik und Mathematik (E)-
Appears in Collections:Publications without fulltext
Show simple item record

Page view(s)

113
Last Week
0
Last month
2
checked on Jan 31, 2023

Google ScholarTM

Check

Add Files to Item

Note about this record

Export

Items in TORE are protected by copyright, with all rights reserved, unless otherwise indicated.