Pascal Lorenz’s research while affiliated with University of Upper Alsace and other places

Unmanned Aerial Vehicles (UAVs) are increasingly recognized for their pivotal role in military and civilian applications, serving as essential technology for transmitting, evaluating, and gathering information. Unfortunately, this crucial process often occurs through unsecured wireless connections, exposing it to numerous cyber-physical attacks. Furthermore, UAVs' limited onboard computing resources make it challenging to perform complex cryptographic operations. The main aim of constructing a cryptographic scheme is to provide substantial security while reducing the computation and communication costs. This article introduces a very efficient and secure cross-domain Authenticated Key Agreement (AKA) scheme that uses Hyperelliptic Curve Cryptography (HECC). The HECC, a modified version of Elliptic Curve Cryptography (ECC) with smaller parameters and a maximum key size of 80 bits, is well-suited for use in UAVs. In addition, the proposed scheme is employed in a cross-domain environment that integrates a Public Key Infrastructure (PKI) at the receiving end and a Certificateless Cryptosystem (CLC) at the sending end. Integrating CLC with PKI improves network security by restricting the exposure of encryption keys only to the message's sender and subsequent receiver. A security study employing ROM and ROR models, together with a comparative performance analysis, shows that the proposed scheme outperforms comparable existing schemes in terms of both efficiency and security.

The Internet of Things (IoT) and its industrial counterpart, the Industrial Internet of Things (IIoT), have transformed sectors such as home automation, healthcare, and manufacturing by enhancing data management through advanced networking. However, the rapid growth of IIoT has introduced significant cybersecurity challenges, necessitating a comprehensive approach to securing data across the TCP/IP model. This paper presents a novel cybersecurity investment strategy formulated as a bi-objective optimization problem, validated through genetic and iterative algorithms. The strategy effectively balances security and cost, achieving nearly 50% efficiency in solution effectiveness. By utilizing these optimization techniques, the approach provides a practical and cost-effective solution to improve IIoT security within budget constraints, offering valuable insights for cybersecurity professionals seeking robust and economically viable solutions.

Parkinson’s Disease (PD) is a progressive neurodegenerative disorder affecting millions worldwide. Early detection is crucial for improving patient outcomes. Spiral drawing analysis has emerged as a non-invasive tool to detect early motor impairments associated with PD. This study examines the performance of hybrid deep learning and machine learning models in detecting PD using spiral drawings, with a focus on the impact of data augmentation techniques. We compare the accuracy of Vision Transformer (ViT) with K-Nearest Neighbors (KNN), Convolutional Neural Networks (CNN) with Support Vector Machines (SVM), and Residual Neural Networks (ResNet-50) with Logistic Regression, evaluating their performance on both augmented and non-augmented data. Our findings reveal that ViT with KNN, initially achieving 96.77% accuracy on unaugmented data, experienced a notable decline across all augmentation techniques, suggesting it relies heavily on global patterns in spiral drawings. In contrast, ResNet-50 with Logistic Regression showed consistent improvement with data augmentation, reaching 93.55% accuracy when rotation and flipping techniques were applied. These results highlight that hybrid models respond differently to augmentation, and careful selection of augmentation strategies is necessary for optimizing model performance. Our study provides important insights into the development of reliable diagnostic tools for early PD detection, emphasizing the need for appropriate augmentation techniques in medical image analysis.

Federated Learning (FL) enabled the reliability and robustness of 5G communication networks for wireless edge computing to provide collaborative Deep Learning (DL) of complex models while protecting privacy for healthcare systems. Wireless end devices are more susceptible to corruption due to the vulnerability offered by open network settings, however, this creates security issues and lessens the effectiveness of DL-based security models for healthcare systems. Furthermore, disaster reliability in communication networks has garnered unprecedented attention from governments and companies, particularly during the current COVID-19 pandemic scenario. In this work, a novel reliable personalized Federated Learning-based Customized Inequality-Aware Federated Learning (CusIAFL) technique is proposed for securing color images while communicating with a wireless network. The proposed technique adjusts each data sampling to the local target during optimization using knowledge of client-label availability. The work that is being presented uses a hybrid technique to maintain consistency in the time-series data. and a novel Pix2Pix Generative Adversarial Network (GAN) technique is used to generate realistic images. This novel work is tested on different non-medical and medical images. The experimental results have been evaluated using performance metrics, namely accuracy, entropy, PSNR, HD95, SSIM, and MSE. Furthermore, the accuracy varies from 89 to 93 percent with different datasets outperforming well with existing SOTA techniques. The outcomes demonstrate that the proposed CusIAFL-based scheme is more effective than the State-Of-The-Art (SOTA) models.

Urban air quality, impacted by human-made pollution, impacts health and requires continuous monitoring. MQ sensors are the preferred air quality sensors despite their high energy consumption due to their cost, requiring the use machine learning to classify different types of air. The aim of this paper is to evaluate a monitoring solution with low-cost and low-energy consumption to classify urban and rural air. A single MQ sensor will be used with a network with edge and fog computing to balance the energy consumption. Edge computing was included in the node for feature extraction, and fog computing was applied in the smartphone to classify the data using machine learning. Different sensors and time buffers are compared in order to find the adequate sensor for data generation and time buffer for feature extraction. The results indicate that it has been possible to achieve accuracies of 100% using a single sensor, the MQ2, with time buffers of 45 to 60 measures. With this proposal, it is possible to reduce the energy consumed by data gathering to 25% of the original consumption due to the use of a single sensor, thanks to the reduction in the sensors used in the previous prototype. Moreover, it has been possible to reduce the energy linked to data forwarding by almost 97 % due to using a time buffer.

In recent years, the rise in cyber threats targeting online platforms has led to significant advancements in cyber-security research. This paper offers a comprehensive overview of current trends and persistent challenges in the cybersecurity landscape. The study evaluates the effectiveness of various cyber-security measures, including attack detection and simulation, incident response, and blockchain-based security frameworks. Despite progress, many cybersecurity solutions are still limited by factors such as insufficient datasets, computational inefficiency, and susceptibility to adversarial attacks. By analyzing recent literature and highlighting these ongoing gaps, this paper aims to direct future research toward developing more robust and adaptive cybersecurity solutions, enhancing protection for online platforms and IT systems against evolving cyber threats.

The increase in popularity of vehicles encourages the development of smart cities. With this advancement, vehicular ad-hoc networks, or VANETs, are now frequently utilized for inter-vehicular communication to gather data regarding traffic congestion, vehicle location, speed, and road conditions. Such a public network is open to various security risks. Overall, protecting personal information on VANET is a vital responsibility. The integration of fog computing and VANETs has gained significant importance in recent years, driven by advancements in cloud computing, Internet of Things (IoT) technologies, and intelligent transportation systems. However, ensuring secure communication in fog-based VANETs remains a major challenge. To overcome this challenge, we introduce a novel authenticated key agreement protocol that achieves mutual authentication, generates a secure session key for secret communication, and provides privacy protection without the use of bilinear pairing. We rigorously prove the security of our proposed protocol, which is designed specifically for fog-based VANETs, and has been shown to meet their stringent security requirements. Moreover, we performed formal and informal analysis that shows our proposed protocol is highly efficient,our protocol's computational and communication overhead are lower than those of other relevant protocols by 45.570% and 29.432%, respectively. Finally we use NS-3 simulation to prove that our proposed algorithm is a practical and scalable solution for secure communication in fog-based VANETs.

The recent increase in the prevalence of skin cancer, along with its significant impact on individuals’ lives, has garnered the attention of many researchers in the field of deep learning models, especially following the promising results observed using these models in the medical field. This study aimed to develop a system that can accurately diagnose one of three types of skin cancer: basal cell carcinoma (BCC), melanoma (MEL), and nevi (NV). Additionally, it emphasizes the importance of image quality, as many studies focus on the quantity of images used in deep learning. In this study, transfer learning was employed using the pre-trained VGG-16 model alongside a dataset sourced from Kaggle. Three models were trained while maintaining the same hyperparameters and script to ensure a fair comparison. However, the quantity of data used to train each model was varied to observe specific effects and to hypothesize about the importance of image quality in deep learning models within the medical field. The model with the highest validation score was selected for further testing using a separate test dataset, which the model had not seen before, to evaluate the model’s performance accurately. This work contributes to the existing body of research by demonstrating the critical role of image quality in enhancing diagnostic accuracy, providing a comprehensive evaluation of the VGG-16 model’s performance in skin cancer detection and offering insights that can guide future improvements in the field.