Adrian Silviu Roman

Dr. Adrian Silviu Roman

Department of Electrical Engineering and Information Technology, George Emil Palade University of Medicine, Pharmacy, Science, and Technology of Targu Mures, Târgu-Mureş, Mureş, Romania
orcid

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
18 December 2021 in Electronics
Verified
Journal Article
Open access
PDF file
ARAdrian Silviu Roman
Adrian Silviu Roman
BGBela Genge
Bela Genge
ADAdrian Duka
Adrian Duka
PHPiroska Haller
Piroska Haller
Modern auto-vehicles are built upon a vast collection of sensors that provide large amounts of data processed by dozens of Electronic Control Units (ECUs). These, in turn, monitor and control advanced technological systems providing a large palette of features to the vehicle’s end-users (e.g., automated parking, autonomous vehicles). As modern cars become more and more interconnected with external systems (e.g., cloud-based services), enforcing privacy on data originating from vehicle sensors is becoming a challenging research topic. In contrast, deliberate manipulations of vehicle components, known as tampering, require careful (and remote) monitoring of the vehicle via data transmissions and processing. In this context, this paper documents an efficient methodology for data privacy protection, which can be integrated into modern vehicles. The approach leverages the Fast Fourier Transform (FFT) as a core data transformation algorithm, accompanied by filters and additional transformations. The methodology is seconded by a Random Forest-based regression technique enriched with further statistical analysis for tampering detection in the case of anonymized data. Experimental results, conducted on a data set collected from the On-Board Diagnostics (OBD II) port of a 2015 EUR6 Skoda Rapid 1.2 L TSI passenger vehicle, demonstrate that the restored time-domain data preserves the characteristics required by additional processing algorithms (e.g., tampering detection), showing at the same time an adjustable level of privacy. Moreover, tampering detection is shown to be 100% effective in certain scenarios, even in the context of anonymized data.

7 March 2023 in Applied Sciences
Verified
Journal Article
Open access
PDF file
BGBela Genge
Bela Genge
PHPiroska Haller
Piroska Haller
ARAdrian Silviu Roman
Adrian Silviu Roman
Advanced Persistent Threats (APTs) represent a complex series of techniques directed against a particular organization, where the perpetrator is able to hide its presence for a longer period of time (e.g., months, years). Previous such attacks have demonstrated the exceptional impact that a cyber attack may have on the operation of Supervisory Control And Data Acquisition Systems (SCADA), and, more specifically, on the underlying physical process. Existing techniques for the detection of APTs focus on aggregating results originating from a collection of anomaly detection agents. However, such approaches may require an extensive time period in case the process is in a steady-state. Conversely, this paper documents E-APTDetect, an approach that uses dynamic attestation and multi-level data fusion for the early detection of APTs. The methodology leverages sensitivity analysis and Dempster-Shafer’s Theory of Evidence as its building blocks. Extensive experiments are performed on a realistic Vinyl Acetate Monomer (VAM) process model. The model contains standard chemical unit operations and typical industrial characteristics, which make it suitable for a large variety of experiments. The experimental results conducted on the VAM process demonstrate E-APTDetect’s ability to efficiently detect APTs, but also highlight key aspects related to the attacker’s advantage. The experiments also highlight that the adversary’s advantage is affected by two major factors: the number of compromised components; and, the precision of manipulation.

5 March 2023 in Mathematics
Verified
Journal Article
Open access
PDF file
ARAdrian Silviu Roman
Adrian Silviu Roman
Data collected from sensor-rich systems may reveal user-related patterns that represent private information. Sensitive patterns from time-series data can be protected using diverse perturbation methods; however, choosing the method that provides the desired privacy and utility level is challenging. This paper proposes a new procedure for evaluating the utility and privacy of perturbation techniques and an algorithm for comparing perturbation methods. The contribution is significant for those involved in protecting time-series data collected from various sensors as the approach is sensor-type-independent, algorithm-independent, and data-independent. The analysis of the impact of data integrity attacks on the perturbed data follows the methodology. Experimental results obtained using actual data collected from a VW Passat vehicle via the OBD-II port demonstrate the applicability of the approach to measuring the utility and privacy of perturbation algorithms. Moreover, important benefits have been identified: the proposed approach measures both privacy and utility, various distortion and perturbation methods can be compared (no matter how different), and an evaluation of the impact of data integrity attacks on perturbed data is possible.

12 September 2025 in IEEE Access
Verified
Journal Article
Open access
ARAdrian Silviu Roman
Adrian Silviu Roman
Roland BolboacaRoland Bolboaca
Roland Bolboaca
TLTeri Lenard
Teri Lenard
PHPiroska Haller
Piroska Haller
Location-Based Service (LBS) are increasingly integrated into our daily life, to offer smart services and applications. However, ensuring user privacy while maintaining data utility remains a critical challenge. Trajectory data, vital for urban planning, mobility analytics or, is particularly sensitive to re-identification even when explicit identifiers are removed. This article introduces APU-TrajGen+, a novel method for real-time, privacy-preserving point-to-point trajectory generation. APU-TrajGen+ leverages a Gated Recurrent Unit-based model to synthesize trajectory points, followed by an adaptive perturbation mechanism that dynamically balances privacy and utility. Designed for local execution, APU-TrajGen+ protects location data at the device level, eliminating the need for centralized processing and preventing unprotected data transmission. Unlike many existing methods, APU-TrajGen+ does not rely on map data or points of interest to generate points in plausible locations. A configurable utility-privacy score guides an adaptive process, to allow data owners to set the protection levels to their specific needs. Experimental results show that APU-TrajGen+ produces realistic trajectories while achieving desired privacy and utility objectives. The method is particularly suitable for LBS and TMS scenarios that require continuous, point-to-point data release, supporting both high utility and strong privacy guarantees without compromising real-time performance.

26 October 2023
Verified
Conference Paper
Closed access
ARAdrian Silviu Roman
Adrian Silviu Roman
BGBela Genge
Bela Genge
The proliferation of WiFi access points has significantly transformed our interactions with the digital world. However, the crowdsourced collection and dissemination of WiFi access point data have raised considerable concerns regarding privacy. The availability of such data can facilitate the identification of access points, and the presence of identifiable information, including names, can enable inferences regarding the location of an individual’s residence, thereby jeopardizing the privacy of sensitive personal information. In response to these concerns, we propose a novel approach to privacy-preserving crowdsourced WiFi data publishing, utilizing differential privacy. This approach ensures that the released data maintains the confidentiality of WiFi access point owners while enabling useful analysis of the network data. The objective of this article is to investigate the challenges and possibilities associated with the use of differential privacy in crowdsourced RSS data publishing for WiFi access points, with an emphasis on preserving privacy. The proposed method is evaluated in a simulated environment using Mininet-WiFi.