RDFdL: Integrating RDF with Differential Dynamic Logic
This paper introduces RDFdL, a framework that integrates RDF knowledge graphs with Differential Dynamic Logic to enable the representation and formal verification of both static knowledge and continuous physical dynamics, thereby allowing safety and reachability properties of cyber-physical systems to be queried via SPARQL.
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
In the modern factory, a digital twin acts as a virtual mirror for a physical machine. It is a detailed map of connections, showing which sensor belongs to which valve, which technician is responsible for a specific motor, and how a product moves from one station to the next. This map is built using a language called the Semantic Web, which excels at organizing static facts and relationships. It can tell an engineer that a heating oven has an "on" switch and an "off" switch. However, this digital map has a blind spot. It cannot predict what happens when the machine actually runs. It cannot calculate whether the temperature inside that oven will rise too fast, whether a tank will overflow, or if a robot arm will collide with a wall while moving. These questions depend on the laws of physics, specifically how things change over time according to differential equations, which are mathematical rules describing continuous motion. Traditional digital maps are silent on these dynamic behaviors, leaving a dangerous gap between knowing what a machine is and knowing what it will do.
Researchers at the Karlsruhe Institute of Technology have built a bridge to close this gap. They developed a new framework called RDFdL, which fuses the static knowledge of digital maps with the rigorous logic used to verify complex physical systems. The team took the standard language used to describe factory data and taught it to speak the language of dynamic physics. They created a system where the computer can not only read a list of machine parts but also run a formal proof to determine if a specific sequence of events is safe. In their approach, the computer translates a description of a machine's state—such as "the oven is on and the temperature is below 180 degrees"—into a mathematical model. It then uses a specialized theorem prover, a tool designed to check the truth of complex logical statements, to verify if the machine can safely move from that state to a new one, like "the oven is on and the temperature is between 180 and 200 degrees."
The researchers tested this system on several real-world scenarios, including a simple oven, a system of two water tanks, a yogurt production line, and a large industrial steam boiler. For each case, they first described the machine and its possible states using standard digital map tools. They then asked the system to verify if the machine could physically transition between these states without breaking safety rules, such as exceeding a maximum temperature or pressure. The system successfully proved that for the oven, it is possible to heat up from a cold state to a target range without ever surpassing the safety limit of 200 degrees. It confirmed that the water in the tanks would flow correctly and that the steam boiler could maintain safe pressure levels. Crucially, once the system proved a transition was safe, it added that fact back into the digital map as a verified connection. This means that a manager can now ask a single question in a standard query language: "Can the machine go from state A to state B, and if so, which technician is responsible for the device involved?" The system answers by combining the static map of personnel and parts with the newly verified path of physical behavior.
The work demonstrates that it is possible to integrate the continuous, flowing nature of physical laws with the structured, discrete nature of data graphs. The team showed that their method works for systems with linear behaviors, where changes happen at a steady rate, and also for more complex, non-linear systems where variables interact in difficult ways, such as the flow of water between two tanks. They found that the number of checks required grew in a manageable way as the systems became larger, suggesting the approach could scale to industrial use. The researchers did not claim to have solved every problem in factory automation, but they proved that the two worlds of static data and dynamic physics can be linked. By turning the results of complex mathematical proofs into simple, queryable facts, they have given digital twins the ability to reason about the future safety of the machines they represent, ensuring that the virtual map reflects not just what the machine is, but what it can safely become.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.