AI REDGIO 5.0 Smart Data Enabler
AI REDGIO 5.0 Smart Data Enabler
How an organization relying in Manufacturing, with limited IT expertise, can easily gain insights from its process data.
Asset Objectives
Smart Data Enabler (SDE) is a basic stack consisting in an AI Edge-to-Cloud minimum viable application using AI REDGIO 5.0 recommended Open-Source tools. SDE application goal is to demonstrate how a SME can easily run a short number of simple steps in order to conduct an initial exploration and analysis of their (edge) production data.
Asset Description
SDE is a technological demo showing how a very simple architecture based on open-source tools recommended in AI REDGIO 5.0 project allows SMEs to obtain useful analytical insights from their process data even in conditions of limited maturity and complexity, with reduced technological expertise and practically zero costs/investments. Once SDE has been easily installed locally, the company only needs to adapt the pipeline provided to its usage scenario, launch it, and start exploring its data in search of interesting behaviors.
Production data can be a JSON file like this:
{
"operation": "shift",
"spec": {
"deviceId": "sensor-001",
"timestamp": "2025-12-10 10:56:02",
"readings": {
"temperature": "12.4",
"humidity": "4.2",
"pressure": "18.5",
"battery": "-1.12"
},
"status": "OK"
}
}
The application stack is composed by the following open-source technologies, whose (integrated) usage is recommended by AIREDGIO 5.0 project:
- Apache NiFi: easy to use, powerful, and reliable open-source system to process and distribute data, particularly suitable for integrating IoT data sources
- MinIO: high-performance, software-defined Object Storage server, a sort of an open-source, private version of Amazon S3
- InfluxDB: specialized open-source database designed to handle data that is indexed by time, fitting best when capturing streams of measurements coming from a sensors
- Grafana: open-source visualization and analytics platform, allowing you to query, visualize, alert on, and understand your metrics no matter where they are stored
and this is how the scenario turns into architecture, leveraging the previous technologies:
Nevertheless, the architecture is ready for various types of improvements thanks to AI.
Use cases
SDE application consists of a base stack and four ready-to-use industrial case studies on top:
- Electrical panel monitoring (small data) Four IIoT sensors – one for each phase R, S, T and Neutral - positioned inside a critical high-voltage substation wide electrical panel, a cooled environment for which it is fundamental to check data in order to highlight anomalies. A relevant objective is to detect whether one of the terminals is loosening (causing an increase in resistance and therefore heat) before an electric arc is triggered.
- Electrical panel monitoring (large data) The same use case, with the difference that the data is not small and directly entered into the pipeline, but rather large and read from an external file.
- Data Center environmental monitoring Monitoring of a data center environmental measures expressed in SenML standard. In this use case, we simulate a reading every 60 seconds for a control unit that monitors temperature and humidity.
- Robotic arm telemetry Monitoring a robotic arm activity.
In an industrial robotic arm (e.g., an anthropomorphic robot on an assembly line), telemetry monitors not only the environment but also the mechanical and electrical status of individual joints (axes). Critical variables usually concern the position, current consumption (which indicates stress), and temperature of the motors.