Optimizing LoRa for Edge Computing with TinyML Pipeline for Channel Hopping

TL;DR

This work tackles the problem of reliable, scalable LoRa communication within an IoT–edge–cloud compute continuum by proposing a TinyML-based channel hopping strategy. It formulates a multi-objective framework with the channel-usage variable $x_{i,g,f}(\\tau)$ and collision objectives $\\mathcal{O}_1$ and $\\mathcal{O}_2$, and implements a practical TinyML pipeline that trains on the edge and performs on-device inference for end-nodes, with OTA updates. An open-source reference architecture and a plant recommender case study demonstrate end-to-end integration of LoRa, edge inference, and cloud-based collaborative filtering via cosine similarity, achieving substantial improvements in RSSI ($\\approx 63\%$) and SNR ($\\approx 44\%$) over random hopping. The results validate the feasibility of open-source, low-power LoRa deployments in a compute continuum for urban microfarming, and lay the groundwork for broader IoT deployments with TinyML-driven spectrum management.

Abstract

We propose to integrate long-distance LongRange (LoRa) communication solution for sending the data from IoT to the edge computing system, by taking advantage of its unlicensed nature and the potential for open source implementations that are common in edge computing. We propose a channel hoping optimization model and apply TinyML-based channel hoping model based for LoRa transmissions, as well as experimentally study a fast predictive algorithm to find free channels between edge and IoT devices. In the open source experimental setup that includes LoRa, TinyML and IoT-edge-cloud continuum, we integrate a novel application workflow and cloud-friendly protocol solutions in a case study of plant recommender application that combines concepts of microfarming and urban computing. In a LoRa-optimized edge computing setup, we engineer the application workflow, and apply collaborative filtering and various machine learning algorithms on application data collected to identify and recommend the planting schedule for a specific microfarm in an urban area. In the LoRa experiments, we measure the occurrence of packet loss, RSSI, and SNR, using a random channel hoping scheme to compare with our proposed TinyML method. The results show that it is feasible to use TinyML in microcontrollers for channel hopping, while proving the effectiveness of TinyML in learning to predict the best channel to select for LoRa transmission, and by improving the RSSI by up to 63 %, SNR by up to 44 % in comparison with a random hopping mechanism.

Figures (14)

• Figure 1: LoRa for Edge Computing in a Compute Continuum.
• Figure 2: TinyML flowchart
• Figure 3: Placement of TinyML in IoT LoRa end-nodes
• Figure 4: PLacement of TinyML on LoRa Gateway nodes (GW)
• Figure 5: Engineering the application workflow, inspired by the photo from our visit in Medellín, Colombia.