Environment mapping using wireless Channel State Information and deep learning

Adrian Donarski, Iain Collings, Stephen Hanly

Research output: Chapter in Book/Report/Conference proceedingConference proceeding contributionpeer-review

Abstract

This paper uses wireless Channel State Information (CSI) and deep learning for obtaining situational awareness through environment mapping. Algorithms are developed that permit estimation of interior room dimensions using wireless Single-Input Single-Output (SISO) Channel Impulse Responses (CIRs). The impact of signal received power, and signal bandwidth, on the ability to predict room dimensions is further investigated. Under such CIR limitations, room dimension estimation accuracy of greater than 90% is achieved using Recurrent Neural Networks (RNNs) with a maximum predictive error of 0.5m. This demonstrates that existing consumer wireless systems with similar physical constraints, may be capable of accurately estimating room dimensions given multiple CIR measurements from multiple receiver locations.

Original languageEnglish
Title of host publication2020, 14th International Conference on Signal Processing and Communication Systems, (ICSPCS)
Subtitle of host publicationproceedings
EditorsTadeusz A. Wysocki, Beata J. Wysocki
Place of PublicationPiscataway, NJ
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Pages1-9
Number of pages9
ISBN (Electronic)9781728199726, 9781728199719
ISBN (Print)9781728199733
DOIs
Publication statusPublished - 2020
Event14th International Conference on Signal Processing and Communication Systems, ICSPCS 2020 - Virtual
Duration: 14 Dec 202016 Dec 2020

Conference

Conference14th International Conference on Signal Processing and Communication Systems, ICSPCS 2020
CityVirtual
Period14/12/2016/12/20

Fingerprint Dive into the research topics of 'Environment mapping using wireless Channel State Information and deep learning'. Together they form a unique fingerprint.

Cite this