Publications
-
Synergistic data-resource participant selection for efficient Federated Edge Learning in IoT ecosystems
The Internet of Things (IoT), as a concept, is becoming increasingly integral to our daily lives, enabling smart environments through sensing, communication, and computation. However, real-world edge devices exhibit pronounced heterogeneity and inherent limitations in both computational resources and data distributions, posing significant challenges for deploying robust, efficient, and adaptive edge intelligence. We propose FedCDRP, […]
-
Decision-support analytics for material selection for production tooling: A systematic review and multi-objective optimisation of biocomposites
Biocomposites are increasingly promoted as sustainable substitutes for conventional composites, yet material selection remains uncertain because existing studies rely on heterogeneous datasets, inconsistent sustainability indicators, and divergent decision models. This paper addresses this gap by combining a systematic literature review with an auditable decision-support workflow for sustainable biocomposite selection, retaining 58 primary studies. The synthesis […]
-
DRIFT: Distributed RPL Injection Flooding Threat in IoT Networks
The Routing Protocol for Low-Power and Lossy Networks (RPL) is widely adopted in IoT environments due to its efficiency in handling constrained devices and networks. However, RPL’s vulnerability to specific attacks presents significant security challenges. This paper introduces DRIFT, a novel distributed DIS attack targeting RPL-based IoT networks. DRIFT attack involves multiple malicious nodes flooding […]
-
Fed-DCSRW: a privacy-preserving, dynamic client selection framework for heterogeneous federated learning via roulette wheel mechanism
We introduce Fed-DCSRW, a federated-learning framework that combines three pillars: (i) a decentralized, data-parallel client-clustering stage that computes centroids in parallel to scale and compute efficiently across heterogeneous clients; (ii) a centroid-based, noise-tolerant Roulette-Wheel client-selection strategy; and (iii) end-to-end differential privacy on both loss reports and gradient updates. The parallel clustering phase partitions the distance […]
-
A Model Predictive Control Approach To Blending In Shared Control
Shared control aims at assisting human operators using robots in physically and cognitively demanding tasks which cannot be automated as they require human expertise and deliberative abilities. Sharing control for a given task typically involves blending algorithms that combine human control inputs and (pre)planned assistance trajectories. Conventional blending techniques, such as Linear Blending, compute a […]
-
Modeling occupant air-conditioning behavior in Mediterranean offices with manual HVAC control: a case study conducted in Montpellier during summer 2025
Occupant behavior related to air conditioning (AC) has a considerable impact on the energy consumption of office buildings, particularly in southern France, which is characterized by a hot Mediterranean climate and frequent heat waves. Although several studies have developed behavioral models using logistic regression, relatively few have examined the application of machine learning methods to […]
-
Predictive maintenance in cyber-physical systems: a comprehensive review of applications, approaches, and challenges
The rapid proliferation of predictive maintenance (PdM) techniques has led to a fragmented research landscape, limiting knowledge consolidation and solution reuse across industrial contexts. Motivated by this issue, this study presents a systematic review of PdM approaches for cyber-physical systems, covering data-driven, statistical, stochastic, AI-based, knowledge-based, and hybrid methodologies. In the context of Industry 4.0 […]
-
Mixed integer quadratic programming for modular robots self reconfiguration problem
Recent years have shown continuous progress in nanoscale assembly with the development of affordable large-scale manufacturing technologies. Modular robotics and, especially, the programmable matter is one of the emanations of this evolution, which consists of a set of identical modular robots capable of reconfiguring into any shape. However, the success of such a technology depends […]
-
Analytical solution of the Langmuir model for moisture diffusion in cylindrical coordinates
Moisture diffusion in polymer composites and bio-based materials frequently exhibits non-Fickian behaviors, such as delayed saturation or two-stage sorption kinetics. While the Langmuir model accurately captures these phenomena by accounting for the dynamic trapping of mobile water molecules, exact analytical solutions have historically been restricted to planar geometries. In this study, the coupled partial differential […]
-
Clustering variables: current status and recent applications
Si les méthodes de clustering d’unités statistiques sont nombreuses, il existe en revanche peu de méthodes spécifiques pour le clustering de variables. Nous présenterons un panorama des méthodes utilisables (divisives, agglomératives, mélanges) en mettant l’accent sur le cas d’ensembles combinant variables quantitatives et qualitatives. Ces différentes approches seront comparées sur des ensembles de données réels, […]
-
Enhanced Energy Delivery in Electric Vehicle Charging Scheduling via Metaheuristic Approaches
Electric vehicle (EV) adoption is rapidly increasing, intensifying pressure on existing charging infrastructures and making the allocation of limited station capacity a critical scheduling problem. In the EV charging scheduling problem, heuristic and hybrid metaheuristic approaches aim to minimize non-delivered energy using discrete-time formulations where each vehicle-charger pair occupies an integer number of time slots. […]
-
BootBoGs-TS: Bootstrap-Based Hyperparameter Optimization for LSTM Models in Remaining Useful Life Prediction
Neural networks, particularly Long Short-Term Memory (LSTM) models, have become prominent for Remaining Useful Life (RUL) prediction due to their ability to capture temporal degradation patterns from multivariate sensor data. However, their predictive performance is highly sensitive to hyperparameter configurations, and conventional optimization methods such as Grid Search or Bayesian Optimization often become inefficient or […]