Regression Analysis of SNR and RSSI Relation in an Indoor LoRaWAN Network with MATLAB Curve Fitting Method

Ömer Yildiz; Sait Eser Karlık

Article Details

Regression Analysis of SNR and RSSI Relation in an Indoor LoRaWAN Network with MATLAB Curve Fitting Method

Authors:

Ömer Yildiz | Sait Eser Karlık

Publisher:

Giresun University

Main Topics:

Elektronik Cihaz ve Sistem Performansı Değerlendirme | Test ve Simülasyon | Elektronik | Sensörler ve Dijital Donanım (Diğer)

Journal:

Karadeniz Fen Bilimleri Dergisi Vol.15 Issue 2, pp.702-717

Document Type:

Article

Publication Date:

6/2025

Document Language:

Turkish

Document ID (Org No.):

2001350564

pISSN:

2564-7377

Keywords:

Abstract:

 Today, cities are growing rapidly. While increasing population reveals the requirement for more efficient use of resources in urban centers; this situation has brought along concepts of smart city and internet of things. Internet of things applications, which enable resource management, tracking, performance analysis and optimization with low-cost, wireless and remote data exchange in living spaces, bring us LoRaWAN. LoRaWAN technology has become widespread thanks to its low infrastructure cost, low battery consumption, easy installation and expandable network structure in long haul communication. In LoRaWAN network planning, RSSI, SNR and SF values are important for network efficiency. In this study, the relation between SNR and RSSI is analyzed using data obtained from an indoor LoRaWAN network and MATLAB curve fitting method. In curve fitting, polynomial, Fourier and Gaussian regression models are analyzed. In terms of goodness-of-fit parameters, it is observed that the equation of the curve with the highest fit to the network data can be obtained with a 9th order polynomial. When the goodness-of-fit comparison of curves fitted to the network data in order to define the relation between SNR and RSSI for the case where the SF parameter is constant at SF7 is made, the R2 value is determined as 0.9362 when polynomial regression model is used, as 0.8459 when Fourier regression model is used, and as 0.8572 when Gaussian regression model is used. RMSE values are found as 0.2440 in polynomial regression model, 0.3895 in Fourier regression model and 0.3828 in Gaussian regression model.

 Today, cities are growing rapidly. While increasing population reveals the requirement for more efficient use of resources in urban centers; this situation has brought along concepts of smart city and internet of things. Internet of things applications, which enable resource management, tracking, performance analysis and optimization with low-cost, wireless and remote data exchange in living spaces, bring us LoRaWAN. LoRaWAN technology has become widespread thanks to its low infrastructure cost, low battery consumption, easy installation and expandable network structure in long haul communication. In LoRaWAN network planning, RSSI, SNR and SF values are important for network efficiency. In this study, the relation between SNR and RSSI is analyzed using data obtained from an indoor LoRaWAN network and MATLAB curve fitting method. In curve fitting, polynomial, Fourier and Gaussian regression models are analyzed. In terms of goodness-of-fit parameters, it is observed that the equation of the curve with the highest fit to the network data can be obtained with a 9th order polynomial. When the goodness-of-fit comparison of curves fitted to the network data in order to define the relation between SNR and RSSI for the case where the SF parameter is constant at SF7 is made, the R2 value is determined as 0.9362 when polynomial regression model is used, as 0.8459 when Fourier regression model is used, and as 0.8572 when Gaussian regression model is used. RMSE values are found as 0.2440 in polynomial regression model, 0.3895 in Fourier regression model and 0.3828 in Gaussian regression model.

 Today, cities are growing rapidly. While increasing population reveals the requirement for more efficient use of resources in urban centers; this situation has brought along concepts of smart city and internet of things. Internet of things applications, which enable resource management, tracking, performance analysis and optimization with low-cost, wireless and remote data exchange in living spaces, bring us LoRaWAN. LoRaWAN technology has become widespread thanks to its low infrastructure cost, low battery consumption, easy installation and expandable network structure in long haul communication. In LoRaWAN network planning, RSSI, SNR and SF values are important for network efficiency. In this study, the relation between SNR and RSSI is analyzed using data obtained from an indoor LoRaWAN network and MATLAB curve fitting method. In curve fitting, polynomial, Fourier and Gaussian regression models are analyzed. In terms of goodness-of-fit parameters, it is observed that the equation of the curve with the highest fit to the network data can be obtained with a 9th order polynomial. When the goodness-of-fit comparison of curves fitted to the network data in order to define the relation between SNR and RSSI for the case where the SF parameter is constant at SF7 is made, the R2 value is determined as 0.9362 when polynomial regression model is used, as 0.8459 when Fourier regression model is used, and as 0.8572 when Gaussian regression model is used. RMSE values are found as 0.2440 in polynomial regression model, 0.3895 in Fourier regression model and 0.3828 in Gaussian regression model.

 Günümüzde şehirler hızla büyümektedir. Artan nüfus, kent merkezlerindeki kaynakların daha verimli kullanılması gerekliliğini ortaya çıkarırken; bu durum akıllı şehir ve nesnelerin interneti kavramlarını beraberinde getirmiştir. Yaşam alanlarında düşük maliyetli, kablosuz ve uzaktan veri alışverişi ile kaynak yönetimi, takibi, performans analizi ve optimizasyonu yapmaya imkân sağlayan nesnelerin interneti uygulamaları, LoRaWANı karşımıza çıkarmaktadır. LoRaWAN teknolojisi, uzak mesafe iletişimde düşük altyapı maliyeti, düşük pil tüketimi, kolay kurulum ve genişleyebilir ağ yapısı özelliği sayesinde yaygınlaşmıştır. LoRaWAN ağ planlamasında, RSSI, SNR ve SF değerleri ağ verimliliği için önem arz eder. Bu çalışmada, bir bina içi LoRaWAN ağından elde edilen veriler ve MATLAB eğri uydurma yöntemi kullanılarak SNR ile RSSI ilişkisinin analizi gerçekleştirilmiştir. Eğri uydurmada, polinom, Fourier ve Gauss regresyon modelleri incelenmiştir. Uyum iyiliği parametreleri açısından, ağ verilerine en yüksek uyumu sağlayan eğrinin denkleminin 9. dereceden bir polinom ile ifade edilebildiği gözlenmiştir. SF parametresinin SF7 değerinde sabit olduğu durumda, SNR ile RSSI arasındaki ilişkiyi tanımlamak için ağ verilerine uydurulan eğrilerin uyum iyiliği karşılaştırması yapıldığında; R2 değeri, polinom regresyon modeli kullanıldığında 0.9362, Fourier regresyon modeli kullanıldığında 0.8459, Gauss regresyon modeli kullanıldığında ise 0.8572 olarak belirlenmiştir. RMSE değerleri ise, polinom regresyon modelinde 0.2440, Fourier regresyon modelinde 0.3895 ve Gauss regresyon modelinde 0.3828 olarak bulunmuştur.

Article Details

Regression Analysis of SNR and RSSI Relation in an Indoor LoRaWAN Network with MATLAB Curve Fitting Method

Authors:

Ömer Yildiz | Sait Eser Karlık

Publisher:

Giresun University

Main Topics:

Elektronik Cihaz ve Sistem Performansı Değerlendirme | Test ve Simülasyon | Elektronik | Sensörler ve Dijital Donanım (Diğer)

Journal:

Karadeniz Fen Bilimleri Dergisi Vol.15 Issue 2, pp.702-717

Document Type:

Article

Publication Date:

6/2025

Document Language:

Turkish

Document ID (Org No.):

2001350564

pISSN:

2564-7377

LoRaWAN

RSSI

SNR

SF

Curve fittingLoRaWAN

RSSI

SNR

SF

Eğri uydurma

Issues

Article Details

Authors: Ömer Yildiz | Sait Eser Karlık

Publisher: Giresun University

Main Topics: Elektronik Cihaz ve Sistem Performansı Değerlendirme | Test ve Simülasyon | Elektronik | Sensörler ve Dijital Donanım (Diğer)

Journal: Karadeniz Fen Bilimleri Dergisi Vol.15 Issue 2, pp.702-717

Document Type: Article

Publication Date: 6/2025

Document Language: Turkish

Document ID (Org No.): 2001350564

pISSN: 2564-7377

Issues

Authors:

Ömer Yildiz | Sait Eser Karlık

Publisher:

Giresun University

Main Topics:

Elektronik Cihaz ve Sistem Performansı Değerlendirme | Test ve Simülasyon | Elektronik | Sensörler ve Dijital Donanım (Diğer)

Journal:

Karadeniz Fen Bilimleri Dergisi Vol.15 Issue 2, pp.702-717

Document Type:

Article

Publication Date:

6/2025

Document Language:

Turkish

Document ID (Org No.):

2001350564

pISSN:

2564-7377