LoRa Signal Demodulation Using Deep Learning, a Time-Domain Approach

The LoRa modulation scheme is becoming one of the most adopted Internet of Things wireless physical layer due to its ability to transmit data over long distances with low power requirements. Typical demodulation techniques for LoRa utilize variants of non-coherent Frequency Shift Keying demodulation. This paper aims to capitalize on the robustness of deep learning techniques, specifically by using convolutional neural networks to demodulate LoRa symbols. We achieve this by building a dataset consisting of emulated time-domain LoRa symbols across a range of channel impairments; namely, we examine additive white Gaussian noise together with carrier frequency offset and time offset. The presented results show an improvement when utilizing deep learning over typical non-coherent detection while performing very close to the optimal matched filter.

I. Introduction

With the rise of the Internet of Things (IoT), business frameworks relying on objects wirelessly connected over the Internet are currently gaining extensive momentum [1]. Low power wide area networks (LPWAN) technologies focus on long-distance communication applications with energy constraints, which usually comes at the cost of a reduced bit rate. Many of these technologies operate in the license-free bands, posing an additional challenge because of the elevated interference levels [2]-[5]. To ensure reliability, efficient signal demodulation techniques are paramount to enable long-distance communications and better service quality. Thus, reducing the need for signal re-transmission, which entails congesting the scarce spectrum resources. Spectrum efficiency is particularly important in the unlicensed industrial, scientific, and medical (ISM) spectrum band [6] since this band is available to the public and requires minimal fees to access [7]. A prominent LPWAN modulation technique is LoRa, providing a suitable platform for many IoT applications such as utility metering, digital farming, smart building, and logistics tracking [8]. LoRa is a derivative of the chirp spread spectrum (CSS) modulation [9], where it utilizes a frequency-modulated pulse to spread the transmission over the channel’s bandwidth.