RSSI-Based LoRa Simulation Environment: An MLP Application for Position Estimation

Abstract

In this study, an interactive simulation framework was developed to analyze the behavior of Received Signal Strength Indicator (RSSI) values in LoRa-based communication links. This Python-based framework was specifically designed to model a two-dimensional space containing fixed LoRa nodes, mobile targets, and multiple obstacles made of materials such as concrete, brick, wood, glass, and vegetation. Signal attenuation was calculated by combining the Free Space Path Loss (FSPL) model with material-based attenuation coefficients. The accuracy and functionality of the developed framework were validated through position estimation models based on Multi-Layer Perceptron (MLP) networks trained with simulated RSSI data. Nine models were constructed across three environmental scenarios and three sensor configurations to systematically evaluate the impact of sensor density and environmental complexity on positioning accuracy. The findings indicate that the proposed method demonstrates robust estimation performance even in environments with dense and complex obstacles. Furthermore, the ability of the system to simultaneously track multiple moving targets demonstrates the adaptability of the developed framework to practical applications. In conclusion, this study provides a flexible and reproducible research platform for evaluating RSSI-based position estimation algorithms and machine learning approaches.