An object-centric core metamodel for IoT-enhanced event logs
This paper addresses the fragmentation in IoT-process data integration by proposing a core object-centric metamodel that synthesizes existing approaches to facilitate data sharing and collaboration in process mining, validated through a Python implementation across various use cases.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine you are trying to tell the story of a busy factory. You have two very different types of storytellers:
- The Process Manager: They keep a neat, organized diary. They write down big events like "Tank Filled," "Product Mixed," or "Bottle Packed." They care about the steps of the recipe.
- The IoT Sensors: These are like thousands of hyper-active, nervous little spies. They don't care about recipes; they just scream out raw data every millisecond: "Pressure is 204!", "Temperature dropped!", "Flow rate spiked!"
The Problem:
For a long time, these two storytellers couldn't talk to each other. The Manager's diary was too high-level, and the Sensors' data was too messy and low-level. Researchers tried to build different "translation dictionaries" to combine them, but everyone built their own dictionary with different rules. One researcher's "dictionary" couldn't read another's notes. It was like trying to have a conversation where one person speaks French, another speaks German, and a third speaks a made-up language. It made sharing data and learning from each other incredibly difficult.
The Solution: The "CORE" Metamodel
This paper introduces a new, universal translator called the CORE metamodel. Think of it as a new, super-flexible "Universal Logbook" that everyone can agree to use.
Here is how it works, using simple analogies:
1. The "Object-Centric" Approach (The Party Guest Analogy)
Old logbooks were like a list of "Cases." If you were tracking a wedding, you had one list for the "Wedding Case." If you wanted to see what the "Cake" was doing, you had to squint at the wedding list.
The CORE model is like a guest list at a party.
- Objects are the guests (e.g., The Cake, The Bride, The Tank, The Robot).
- Events are the things that happen to them (e.g., "Cake was baked," "Tank was filled").
- In this new book, one event can be linked to many guests at once. A "Mixing" event can be linked to the "Tank," the "Batch of Chemicals," and the "Operator" all at the same time. This solves the problem of trying to force everything into a single, rigid box.
2. Handling Different "Granularities" (The Zoom Lens Analogy)
The CORE model is like a camera with a perfect zoom lens.
- Zoomed Out (Process Events): It sees the big picture: "We started the production line."
- Zoomed In (IoT Events): It sees the tiny details: "The sensor saw a vibration at 12:05:03."
- The Magic Link: The model doesn't just store both; it draws a line between them. It says, "This big 'Start Production' event was actually made up of these 50 tiny sensor vibrations." It can even show you the "recipe" (the algorithm) used to turn those tiny vibrations into the big event.
3. The "Data Sources" (The Mailman Analogy)
In the old days, logs often forgot where the data came from. The CORE model treats the Source as a character in the story.
- Did the data come from a Sensor (a physical spy)?
- Did it come from an Information System (a digital clerk)?
- Did it come from a Link (a translator that combined two other pieces of data)?
The model records exactly who delivered the message, so you know if you are looking at a raw temperature reading or a calculated "Quality Score."
How They Tested It
The authors didn't just write a theory; they built a prototype (a working model) in Python. They took three very different, real-world stories and tried to translate them into this new Universal Logbook:
- A Smart Home: Tracking elderly people's daily routines (cooking, sleeping) using sensors.
- A Manufacturing Plant: Tracking chemical mixing and bottling with high-speed sensors.
- A Blood Donation Center: Tracking hand hygiene and medical steps to ensure safety.
The Result:
They successfully translated all three stories into the CORE format without losing any important details. It was like taking a French novel, a German play, and a Japanese haiku, and successfully rewriting them all into a single, coherent English story that kept all the original meaning.
Why This Matters
- For Researchers: It stops everyone from reinventing the wheel. Now, if a scientist in Germany wants to share data with a scientist in Brazil, they can just swap files in this new format. No more translation headaches.
- For Industry: Companies can finally mix their "business steps" (like "Order Shipped") with their "machine data" (like "Conveyor Belt Speed") in one place. This helps them spot problems faster, like realizing that a specific machine vibration always leads to a bad product.
In short, the paper says: "We built a universal language for IoT and business processes so that data from different worlds can finally hold hands and tell a complete story."
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.