Co-simulation is a key step in the development of today’s complex cyber-physical systems (CPS), specially in the integration and validation activities. However, performing a co-simulation involving models developed in different environments and possibly deployed in different platforms with mixed real-time and non real-time constraints is a challenging engineering task. A promising technology that could help overcome communication and synchronisation difficulties is the non-proprietary standard Distributed Co-simulation Protocol (DCP). This standard defines an application-level communication protocol, independent of the platform and the communication medium, that regulates the exchange of information between the co-simulation entities. This paper presents a new approach to apply the DCP standard. Instead of using it as a model encapsulation mechanism, having to develop an specific DCP slave for each application, it is proposed to use it as a generic co-simulation interface. To this end, a Simulink library has been developed, allowing to connect models developed in Simulink with the outside world in an standardised way. Moreover, by exploiting the code generation potential of Simulink, a wide variety of devices become accessible, thus enabling x-in-the-loop simulations, which are commonly used tests in the verification and validation process of CPSs. This library has been tested in a soft real-time co-simulation application between a Simulink instance and an application running on a Xilinx Zynq Ultrascale+ System-on-Chip. As an additional contribution, an analysis of DCP synchronisation problems when simulating closed-loop systems composed of two slaves is performed. Finding that the main causes are the occurrence of random delays and that the simulations of the two slaves start at an arbitrary time. A possible solution to this problem is also presented.

Cyber–physical systems (CPS) integrate diverse elements developed by various vendors, often dispersed geographically, posing significant development challenges. This paper presents an improved version of our previously developed co-simulation interface based on the non-proprietary Distributed Co-Simulation Protocol (DCP) standard, now optimized for broader hardware platform compatibility. The core contributions include a demonstration of the interface’s hardware-agnostic capabilities and its straightforward adaptability across different platforms. Furthermore, we provide a comparative analysis of our interface against the original DCP. It is validated via various X-in-the-Loop simulations, reinforcing the interface’s versatility and applicability in diverse scenarios, such as distributed real-time executions, verification and validation processes, or Intellectual Property protection.

The high cost associated with hardware platforms for implementing Hardware-in-the-Loop (HIL) applications is often prohibitive for many research groups and companies. Therefore, there is an increasing demand to develop cost-effective solutions that can still deliver accurate and reliable results. In this paper, we showcase the practical application of our previously developed co-simulation interface as a cost-effective alternative to replace these expensive hardware setups. Our proposal is composed of a hardware part, consisting of a ZedBoard equipped with commercial ADC and DAC modules, and a software part comprising the developed co-simulation interface. The ZedBoard serves as the primary computation platform and acts as a communication medium with the physical world. On the other hand, the interface allows real-time communication between the ZedBoard and various simulation environments and hardware components via Ethernet. To demonstrate this, we apply the interface in a study of chatter in industrial machining processes. Here, we connect the ZedBoard with Simulink via the interface, while the ADC and DAC modules are wired to the acquisition card of a real-time computing platform. As result, we demonstrate that our co-simulation interface offers an accessible and affordable solution for high-performance real-time simulations.

Large Language Models are increasingly used for high-level robotic reasoning, yet their latency and stochasticity complicate their direct use in low-level control. Moreover, extracting actionable navigation cues from multimodal context incurs inference costs that are challenging for embedded platforms. We present a plug-and-play framework that augments a finite-state machine with asynchronous velocity suggestions generated by a Large Language Model, using an off-the-shelf DistilGPT-2 model running on-device on a Jetson AGX Orin. The system extracts task-relevant cues from the current context and integrates them only if they satisfy deadline, schema, and kinematic validation, thereby preserving a deterministic 50 Hz control loop with a <5 ms fallback path. We compare multiple Large Language Models for embedded robot control and quantify trade-offs among model size, inference time, and output validity. To assess whether the Large Language Models add value beyond signal processing, we include an ablation against a standard smoothing baseline; the results indicate that the Large Language Models contribute anticipatory, context-dependent adjustments that are not captured by filtering alone. Experiments in Gazebo and on a real TurtleBot3 reduce the final position error from 0.246 m to 0.159 m and improve trajectory efficiency from 0.821 to 0.901 without increasing control-loop latency. Approximately 80% of the Large Language Models’ outputs pass validation and are applied. Overall, the framework reduces developer effort by enabling behavioral changes at the prompt level while maintaining interpretable, robust edge-based navigation.

This paper presents a proof-of-concept for a natural-language-based closed-loop controller that regulates the temperature of a simple single-input single-output (SISO) thermal process. The key idea is to express a relay-with-hysteresis policy in plain English and let a local large language model (LLM) interpret sensor readings and output a binary actuation command at each sampling step. Beyond interface convenience, we demonstrate that natural language can serve as a valid medium for modeling physical reality and executing deterministic reasoning in control loops. We implement a compact plant model and compare two controllers: a conventional coded relay and an LLM-driven controller prompted with the same logic and constrained to a single-token output. The workflow integrates schema validation, retries, and a safe fallback, while a stepwise evaluator checks agreement with the baseline. In a long-horizon (1000-step) simulation, the language controller reproduces the hysteresis behavior with matching switching patterns. Furthermore, sensitivity and ablation studies demonstrate the system’s robustness to measurement noise and the LLM’s ability to correctly execute the hysteresis policy, thereby preserving the theoretical robustness inherent to this control law. This work demonstrates that, for slow thermal dynamics, natural-language policies can achieve comparable performance to classical relay systems while providing a transparent, human-readable interface and facilitating rapid iteration.