Nowadays, Cyber-Physical Systems (CPS), particularly drones, play a pivotal role in environmental research. Scientists depend on these platforms to monitor various sensor data and ensure comprehensive data archiving. However, despite their advantages, researchers encounter several challenges, including communication limitations and the complexity of setting up systems tailored to their needs. To address these issues, we propose MoDD, a model-driven data collection framework based on a customized publish/subscribe model. MoDD simplifies the development and configuration of data collection systems. It offers scientists a solution that meets their specific needs, allowing them to focus on high-level requirements while the framework manages the underlying complexities. We demonstrate the effectiveness of MoDD through practical evaluations on an actual Unmanned Surface Vehicle. Additionally, results show a 79% reduction in throughput (drone to base station link) compared to existing publish/subscribe systems.

One primary objective of drone simulation is to evaluate diverse drone configurations and contexts aligned with specific user objectives. The initial challenge for simulator designers involves managing the heterogeneity of drone components, encompassing both software and hardware systems, as well as the drone’s behavior. To facilitate the integration of these diverse models, the Functional Mock-Up Interface (FMI) for co-simulation proposes a generic data-oriented interface. However, an additional challenge lies in simplifying the configuration of co-simulation, necessitating an approach to guide the modeling of parametric features and operational conditions such as failures or environment changes. The article addresses this challenge by introducing CARES, a model-driven engineering and component-based approach for designing drone simulators, integrating the FMI for co-simulation. The proposed models incorporate concepts from component-based software engineering and FMI. The NAVIDRO architectural style is presented for designing and configuring drone co-simulation. CARES utilizes a code generator to produce structural glue code (Java or C++), facilitating the integration of FMI-based domain-specific code. The approach is evaluated through the development of a simulator for navigation functions in an autonomous underwater vehicle, demonstrating its effectiveness in assessing various autonomous underwater vehicle configurations and contexts.

Scientists who analyze physical phenomena usually work on data coming from various sensors spread over the environment. Visualization of these data is a key issue in this process. In addition, these data are massive and provided at a very high rate. Under these conditions, it can be a real challenge to efficiently implement a reliable, dedicated, and itself efficient visualization tool, especially without the support of a large scale platform. In this paper, we propose to take up this challenge with FaST: a model-driven framework which makes it possible to generate a complete solution for the storage, the querying and the visualization of time series in a big data context. The specification of the solution is efficient thanks to a dedicated language intended for data scientists. It enables a simple description of both the architecture of the solution and the data it has to handle. Deployment is efficient thanks to code generation and server-side dockerization. The generated tool is efficient thanks to the implementation of ad hoc optimizations. On server-side, they come from pre-computation of views based on the Min-Max principle. On client-side, they come from the anticipation of queries related to the data navigation abilities of the generated tool. The underlying principles of FaST and its optimizations are detailed in this paper, as well as its implementation and its performance evaluation. It highlights a significant gain in execution time, with a limited database overhead.

One key objective of Cyber-Physical System (CPS) simulation is to evaluate different CPS configurations regarding a certain user objective. First, simulation of CPS necessitates frameworks to handle heterogeneity of CPS components (the software and hardware system control, the behavior of the CPS itself and its physical environment). Then, to build simulators, designers use paradigms like FMI (Functional Mock-Up Interface) that proposes a data-driven generic interface facilitating the integration of heterogeneous models. However, in order to facilitate simulation configuration, an approach is required to drive modeling of parametric features and operational conditions. In this paper, we present CARES, a component-based and model-driven approach to facilitate CPS simulation. CARES is applied to evaluate an Autonomous Underwater Vehicle (AUV) navigation function by simulation. The proposed models integrate both the principles of a generic simulation (integration of Component Based Software Engineering CBSE concepts and FMI paradigm) and domain specific aspects through a component-based architecture style. From a design model, a code generator builds the structural (Java or C++) code of the simulator. The generated code relies on a given run-time library for its execution and its structure facilitates integration of domain-specific code. The experiments show the effectiveness of the approach to build simulators for evaluation of different AUV configurations.

Cyber-physical systems evolving in uncertain environment endure fluctuating dynamics during their lifetime. In such a variable context, controlling systems towards safety and system performances is challenging. In particular, controller tuning (finding optimal control parameters) is a challenging process due to the multiplicity of contexts to be considered. In this paper, we use a combination of model-driven simulation, dimensionality reduction, clustering and prediction techniques to define adequate control parameter settings. First, we propose to explore the controller behavior by simulating different configurations, a configuration is defined by a context (controlled process, environment, sensors, actuators) and a control parameters setting. From simulation results, a discretization is performed by binning the evaluation of quality of control. Then, we apply feature selection algorithms to identify contextual parameters that have a significant impact on performances of the controller. Considering only selected parameters, we finally carry out a clustering aiming at identifying for context domains an optimal control parameter setting. The approach is iterative to define the boundaries of the controller for a given context domain. For non simulated contexts, we propose a prediction module based on regression techniques.To evaluate the proposed approach, we compare it with classical control theory and we apply it to a proportional controller used for a leader/follower application. The experiment shows effectiveness in the identification of control parameters setting for different contexts.

International audience