Project co-funded by the European Union, European Regional Development Fund - National Operational Programme Enterprises and Competitiveness 2014-2020, Axis VI REACT-EU, financed as part of the Union's response to the COVID-19 pandemic.
Funded case
SAMANTA - predictive maintenance for vehicles
SAMANTA (SmArt MAiNTenance plAtform) is a research and development project co-funded by the European Union that built an IoT predictive-maintenance platform for vehicles: on-board units, data collection, a web app and an artificial intelligence module that flags component wear.
Summary
Seeing a vehicle's wear before it becomes a failure.
SAMANTA set out to bring predictive maintenance to vehicles: instead of acting after a failure or on a fixed schedule, the platform collects on-board data, analyses it and flags when a component is wearing out. The project, co-funded by the European Union, brought together Elettronica Tirrito's on-board hardware, Adeo's software platform and the research of Kore University of Enna.
- On-board units with OBD-II diagnostics, environmental sensors and a thermal camera.
- Web platform for real-time data, history and alarms.
- Artificial intelligence module for wear estimation.
Problem
Maintenance on failure or on schedule.
In service fleets, maintenance usually follows the calendar or the failure. In the first case good components are replaced, in the second the vehicle stops suddenly and the repair costs more. The data to do better is already on board, but it is rarely collected and read continuously.
- Emergency interventions and unplanned downtime.
- Diagnostic data available only in the workshop.
- No history to understand how a component wears.
Approach
Continuous data from the vehicle, analysis in the platform.
The project split the work into three layers: on-board acquisition, the platform that receives, stores and shows the data, and the models that interpret it. Adeo led the architecture study, the software implementation and the release of the advanced prototype; Kore University developed the algorithms and the study of the interface for technicians.
- Architecture designed for vehicles with WiFi, 4G or NB-IoT connectivity.
- Requirements, quality plan, integration and testing of the software system.
- Wear-estimation algorithms integrated as platform alarms.
System
From the on-board unit to the web app.
The on-board units read the vehicle diagnostics through OBD-II and the CAN bus, temperature, humidity, accelerations, engine speed and the image of a thermal camera. Data reaches a collection service, is stored in a historical archive and made available through APIs to a web app that shows real-time values, history, the thermal camera heat map and alarms.
- Data collection service for on-board protocols.
- APIs and historical archive of measurements.
- Web app with live, history and heat-map views.
- Study of an augmented reality application for technicians and automatic vehicle recognition.
Results
An advanced prototype tested on vehicles in service.
The project aimed to reduce emergency interventions and repair times through data-driven maintenance. It led to the release of an advanced prototype of the platform, tested on service vehicles such as waste-collection vans, street sweepers and passenger minibuses. The wear-estimation module was integrated as an alarm and verified in simulation; the on-board units with thermal cameras kept sending data until 2025.
- Advanced prototype of the platform released at the end of the project.
- Tests on real vehicles in service.
- Wear alarms integrated in the platform.
- History of on-board data collected in the field.
Technology
The RDM platform applied to vehicles.
SAMANTA's software side is built on Adeo's RDM platform: .NET data collection services, ASP.NET Core APIs, a MongoDB archive and a Blazor web app. The artificial intelligence module was integrated into the same platform, so that a wear estimate becomes an alarm like any other.
- OBD-II diagnostics and CAN bus, environmental and inertial sensors, thermal camera.
- WiFi, 4G and NB-IoT connectivity.
- .NET, ASP.NET Core, MongoDB, Blazor.
- Regression models for wear estimation.
Reuse
From research to platform.
SAMANTA brought into RDM vehicle data collection via OBD-II, thermal camera handling and the integration of predictive models as alarms: capabilities Adeo reuses in later monitoring projects.
- Collect data continuously before training the models.
- Put AI where technicians already look: in the alarms.
- Design the on-board hardware and the platform together.
CTA
Do you have a fleet or machinery to monitor?
Adeo can start from the SAMANTA experience and from RDM to bring predictive maintenance to your vehicles or plants.
Discuss predictive maintenance