Predicting Application Performance in LoRa IoT Networks
Conference proceedings article
Authors/Editors
Strategic Research Themes
Publication Details
Author list: Chungsawat, Natchaya; Siripongwutikorn, Peerapon
Publisher: Hindawi
Publication year: 2020
ISBN: 9781450377591
ISSN: 0146-9428
eISSN: 1745-4557
Languages: English-Great Britain (EN-GB)
Abstract
Network planning is critical to satisfactory application performance and efficient network resource utilization especially in an outdoor IoT network such as LoRa. As IoT devices may locate at different distances from the gateway and use different transmission rates and data generation rates, determining if a given set of devices with heterogenous configurations can be feasibly deployed is a challenging task. Unlike previous works which mostly focus on the effect of transmission techniques in MAC and physical layer parameters, this paper develops prediction models of various IoT application performance metrics in a single-gateway LoRa IoT network given a set of heterogeneous device configurations in terms of distance from gateway, data rate, and packet generation rate, which are more relevant inputs to network provisioning. Performance data of packet loss, average packet delay and high-percentile delays are first obtained from simulation experiments over a wide range of factor values and the data is fitted to binomial regression, linear regression, and neural network models. Our results show that with appropriate model tuning, a standard technique like neural network regression is able to give high prediction accuracy, with prediction errors around 1.5-5.3% on the test dataset depending on the application performance metrics. © 2020 ACM.
Keywords
Application performance, Average delay, High-percentile delays, LoRa
Contact Information