Mobility Classiﬁcation of LoRaWAN Nodes Using Machine Learning at Network Level

Sep 17, 2026

Uploaded by Quinn

LoRaWAN networks rely on an adaptive data rate (ADR) algorithm to maintain link reliability and handle the required density of end devices. Effective operation requires tuning ADR according to each end device’s mobility level. This paper presents an approach using support vector machine–based supervised machine learning to determine mobility level using only data available at the LoRaWAN network server, without requiring location capabilities. Performance in a real network demonstrates reliability and effectiveness of the proposed method.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages9

Reading Time23 min

AccessFree to download

Abstract

Keywords

1. Introduction

sensors Communication Mobility Classiﬁcation of LoRaWAN Nodes Using Machine Learning at Network Level Lorenzo Vangelista 1, *,†, Ivano Calabrese 2,† and Alessandro Cattapan 3,†,‡ Citation: Vangelista, L.; Calabrese, I.; Cattapan, A. Mobility Classiﬁcation of LoRaWAN Nodes Using Machine Learning at Network Level. Sensors 2023, 23, 1806. [https://doi.org/](https://doi.org/)[ ](https://doi.org/)[10.3390/s23041806](10.3390/s23041806) Academic Editor: Carles Gomez Received: 22 December 2022 Revised: 30 January 2023 Accepted: 31 January 2023 Published: 6 February 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ([https://](https://)[ ](https://)[creativecommons.org/licenses/by/](creativecommons.org/licenses/by/)[ ](creativecommons.org/licenses/by/)[4.0/](4.0/)) . 1 Department of Information Engineering, University of Padova, Italy and Wireless and More srl, 35131 Padova, Italy 2 A2ASmartCity, 25124 Brescia, Italy 3 Wireless and More srl, 35131 Padova, Italy * Correspondence: [lorenzo.vangelista@unipd.it](lorenzo.vangelista@unipd.it)[ ](lorenzo.vangelista@unipd.it)† These authors contributed equally to this work. ‡ Current address: Cornèr Banca SA, 6901 Lugano, Switzerland. Abstract: LoRaWAN networks rely heavily on the adaptive data rate algorithm to achieve good link reliability and to support the required density of end devices. However, to be effective the adaptive data rate algorithm needs to be tuned according to the level of mobility of each end device. For that purpose, different adaptive data rate algorithms have been developed for the different levels of mobility of end devices, e.g., for static or mobile end devices. In this paper, we describe and evaluate a new and effective method for determining the level of mobility of end devices based on machine learning techniques and speciﬁcally on the support vector machine supervised learning method. The proposed method does not rely on the location capability of LoRaWAN networks; instead, it relies only on data always available at the LoRaWAN network server. Moreover, the performance of this method in a real LoRaWAN network is assessed; the results give clear evidence of the effectiveness and reliability of the proposed machine learning approach. Keywords: LPWAN; ADR; LoRaWAN 1. Introduction The low power wide area networks (LPWAN) paradigm is gaining a lot of momentum in the ﬁeld of massive Internet of things (IoT) for its peculiarity of providing wide-area coverage while having low power requirements for transmission [1] . In this paper, we focus on LoRaWAN, the most prominent LPWAN technology working in the unlicensed spectrum. LoRaWAN Chapter 3 in Ref. [1] is characterized by a star topology, whereby the LoRa end devices (EDs) are connected to gateways (GWs) that act in principle as simple forwarders toward the network server (NS) . The wireless communication between the EDs and the GWs takes place in the sub-GHz part of the unlicensed spectrum, which is quite limited. Efﬁcient management of the very limited radio resources is thus essential to connect the very large amount of EDs that the LoRaWAN networks are expected to deal with in the massive IoT paradigm. The LoRaWAN networks are often deployed in challenging radio environments in which the variability of the link quality is high due to various factors: obstacles, an urban scenario [2], ED mobility, and deep indoor environments [3] . A key point of the LoRaWAN technology is its ability to trade the data rate for coverage and vice versa, i.e., long-range communication can be established at the cost of a low data rate. The data rate and the power of the EDs are ﬁnely tuned by the adaptive data rate (ADR) algorithm running on LoRaWAN NS. In particular, the canonical LoRaWAN ADR [4] is not very efﬁcient in the case of mob

FAQs

Why does LoRaWAN need mobility-aware adaptive data rate (ADR)?

Because the ADR algorithm must be tuned according to each end device’s mobility level to maintain link reliability and support dense deployments.

What method is proposed to determine the end device mobility level?

The paper proposes a supervised machine learning approach specifically using support vector machine techniques.

Does the proposed method require LoRaWAN location capability?

No. It relies only on data that is always available at the LoRaWAN network server.

Sep 17, 2026

Uploaded by Quinn