LoRa RSSI based outdoor localization in an urban area using random neural networks

Winfred Ingabire; Hadi Larijani; Ryan M. Gibson

LoRa RSSI based outdoor localization in an urban area using random neural networks

Winfred Ingabire^*, Hadi Larijani, Ryan M. Gibson

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

The concept of the Internet of Things (IoT) has led to the interconnection of a significant number of devices and has impacted several applications in smart cities’ development. Localization is widely done using Global Positioning System (GPS). However, with large scale wireless sensor networks, GPS is limited by its high-power consumption and more hardware cost required. An energy-efficient localization system of wireless sensor nodes, especially in outdoor urban environments, is a research challenge with limited investigation. In this paper, an energyefficient end device localization model based on LoRa Received Signal
Strength Indicator (RSSI) is developed using Random Neural Networks (RNN). Various RNN architectures are used to evaluate the proposed model’s performance by applying different learning rates on real RSSI LoRa measurements collected in the urban area of Glasgow City. The proposed model is used to predict the 2D cartesian position coordinates with a minimum mean localization error of 0.39 m.

Original language	English
Title of host publication	SAI Computing Conference 2021
Publisher	Springer
Number of pages	13
Publication status	Accepted/In press - 18 Nov 2020

Publication series

Name	Advances in Intelligent Systems and Computing
Publisher	Springer
ISSN (Print)	2194-5357
ISSN (Electronic)	2194-5365

Keywords

IoT, LoRaWAN, RSSI, localization, RNN

Embargoed Document

LoRa RSSI based outdoor localization in an urban area using random neural networks
Accepted author manuscript, 859 KB

Fingerprint Dive into the research topics of 'LoRa RSSI based outdoor localization in an urban area using random neural networks'. Together they form a unique fingerprint.

Cite this

@inproceedings{31920af85a3f4c32bd2d2de657dcf936,

title = "LoRa RSSI based outdoor localization in an urban area using random neural networks",

abstract = "The concept of the Internet of Things (IoT) has led to the interconnection of a significant number of devices and has impacted several applications in smart cities{\textquoteright} development. Localization is widely done using Global Positioning System (GPS). However, with large scale wireless sensor networks, GPS is limited by its high-power consumption and more hardware cost required. An energy-efficient localization system of wireless sensor nodes, especially in outdoor urban environments, is a research challenge with limited investigation. In this paper, an energyefficient end device localization model based on LoRa Received SignalStrength Indicator (RSSI) is developed using Random Neural Networks (RNN). Various RNN architectures are used to evaluate the proposed model{\textquoteright}s performance by applying different learning rates on real RSSI LoRa measurements collected in the urban area of Glasgow City. The proposed model is used to predict the 2D cartesian position coordinates with a minimum mean localization error of 0.39 m.",

keywords = "IoT, LoRaWAN, RSSI, localization, RNN",

author = "Winfred Ingabire and Hadi Larijani and Gibson, {Ryan M.}",

note = "*1-1-50* Acceptance in SAN. ST 17/02/21 AAM: apply 12m embargo Author confirmed AAM uploaded to Pure is final author version submitted to the conference (email in Conference Papers mailbos sub folder). ST 19/02/21",

year = "2020",

month = nov,

day = "18",

language = "English",

series = "Advances in Intelligent Systems and Computing",

publisher = "Springer",

booktitle = "SAI Computing Conference 2021",

}

TY - GEN

T1 - LoRa RSSI based outdoor localization in an urban area using random neural networks

AU - Ingabire, Winfred

AU - Larijani, Hadi

AU - Gibson, Ryan M.

N1 - *1-1-50* Acceptance in SAN. ST 17/02/21 AAM: apply 12m embargo Author confirmed AAM uploaded to Pure is final author version submitted to the conference (email in Conference Papers mailbos sub folder). ST 19/02/21

PY - 2020/11/18

Y1 - 2020/11/18

N2 - The concept of the Internet of Things (IoT) has led to the interconnection of a significant number of devices and has impacted several applications in smart cities’ development. Localization is widely done using Global Positioning System (GPS). However, with large scale wireless sensor networks, GPS is limited by its high-power consumption and more hardware cost required. An energy-efficient localization system of wireless sensor nodes, especially in outdoor urban environments, is a research challenge with limited investigation. In this paper, an energyefficient end device localization model based on LoRa Received SignalStrength Indicator (RSSI) is developed using Random Neural Networks (RNN). Various RNN architectures are used to evaluate the proposed model’s performance by applying different learning rates on real RSSI LoRa measurements collected in the urban area of Glasgow City. The proposed model is used to predict the 2D cartesian position coordinates with a minimum mean localization error of 0.39 m.

AB - The concept of the Internet of Things (IoT) has led to the interconnection of a significant number of devices and has impacted several applications in smart cities’ development. Localization is widely done using Global Positioning System (GPS). However, with large scale wireless sensor networks, GPS is limited by its high-power consumption and more hardware cost required. An energy-efficient localization system of wireless sensor nodes, especially in outdoor urban environments, is a research challenge with limited investigation. In this paper, an energyefficient end device localization model based on LoRa Received SignalStrength Indicator (RSSI) is developed using Random Neural Networks (RNN). Various RNN architectures are used to evaluate the proposed model’s performance by applying different learning rates on real RSSI LoRa measurements collected in the urban area of Glasgow City. The proposed model is used to predict the 2D cartesian position coordinates with a minimum mean localization error of 0.39 m.

KW - IoT, LoRaWAN, RSSI, localization, RNN

M3 - Conference contribution

T3 - Advances in Intelligent Systems and Computing

BT - SAI Computing Conference 2021

PB - Springer

ER -