Federated Learning emerged as a promising solution to enable collaborative training between organizations while avoiding centralization. However, it remains vulnerable to privacy breaches and attacks that compromise model robustness, such as data and model poisoning. This work presents PRoT-FL, a privacy-preserving and robust Training Manager capable of coordinating different training sessions at the same time. PRoT-FL conducts each training session through a Federated Learning scheme that is resistant to privacy attacks while ensuring robustness. To do so, the model exchange is conducted by a “Private Training Protocol” through secure channels and the protocol is combined with a public blockchain network to provide auditability, integrity and transparency. The original contribution of this work includes: (i) the proposal of a “Private Training Protocol” that breaks the link between a model and its generator, (ii) the integration of this protocol into a complete system, PRoT-FL, which acts as an orchestrator and manages multiple trainings and (iii) a privacy, robustness and performance evaluation. The theoretical analysis shows that PRoT-FL is suitable for a wide range of scenarios, being capable of dealing with multiple privacy attacks while maintaining a flexible selection of methods against attacks that compromise robustness. The experimental results are conducted using three benchmark datasets and compared with traditional Federated Learning using different robust aggregation rules. The results show that those rules still apply to PRoT-FL and that the accuracy of the final model is not degraded while maintaining data privacy.

As the global population continues to grow, the enormous stress on our environment and resources is becoming impossible to ignore. A focus on producing and consuming as cheap as possible has created an economy in which objects are briefly used and then discarded as waste, featuring a linear lifecycle that creates an enormous amount of waste. The alternative to the linear economy "take-make-waste" is called the “circular economy”. Under this paradigm, materials are recycled to build new products or components that are designed and built to promote their reuse and refurbishment. This assures the continuous (re-)exploitation of existing resources, reducing the extraction of new raw materials. However, customers often reject these reused or refurbished products under the suspicion that they do not meet the same usability, safety, or performance levels of new products. In this sense, trustworthy records of historical details of refurbished products could increase consumers’ confidence in products and components of the "circular economy", prioritizing trustworthiness, reliability, and transparency. This work presents a new certification tool based on blockchain technology to guarantee trusted, accurate, transparent and traceable lifecycle information of products and their components and to generate trustworthy certificates to probe refurbished product historical details. This tool aims to enhance refurbished product visibility by creating the basis for making the circular economy a reality in any domain.

The Self-Sovereign Identity (SSI) paradigm has emerged as a decentralized Identity Management model with interesting capabilities for the Internet of Things (IoT). However, current standard SSI protocols and procedures only consider that individuals store their own identity, failing to provide an accurate solution for the identity management of groups where participants might use credentials from different identities and collaborate to meet a set of verifier´s requirements. The present work introduces the concept of Collaborative Credentials (CCs) to formalize identity management procedures that model the collaboration within a group of participants. CCs allow to leverage use cases requiring collaboration that cannot be solved with standard SSI verifiable credentials, increase the privacy of group participants and enable the development of a software framework that any verifier/holder could use to generate a generic application. In the paper, a generic model for CCs is presented, together with an implementation example that is subsequently evaluated in an experimental testbed.

Publisher Copyright: © 2023

Publisher Copyright: © 2021 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved.

The recent increase in reported incidents of security breaches compromising users' privacy call into question the current centralized model in which third-parties collect and control massive amounts of personal data. Blockchain has demonstrated that trusted and auditable computing is possible using a decentralized network of peers accompanied by a public ledger. Furthermore, Homomorphic Encryption (HE) guarantees confidentiality not only on the computation but also on the transmission, and storage processes. The synergy between Blockchain and HE is rapidly increasing in the computing environment. This research proposes a privacy-enhancing distributed and secure protocol for data aggregation backboned by Blockchain and HE technologies. Blockchain acts as a distributed ledger which facilitates efficient data aggregation through a Smart Contract. On the top, HE will be used for data encryption allowing private aggregation operations. The theoretical description, potential applications, a suggested implementation and a performance analysis are presented to validate the proposed solution.

Cognitive medium access control (C-MAC) protocols employ several mechanisms that deal with the heterogeneity of a coexistence scenario with multiple wireless systems in the same band. The underutilized channels of some bands such as the TV white spaces (TVWS), are valuable and tangible resources within the congested spectrum for providing new wireless services on licensed and license-exempt basis. This paper presents a new C-MAC protocol based on a sequential frequency hopping scheme with a proactive spectrum sensing phase for opportunistic spectrum sharing. The proposal has been implemented over a software defined radio (SDR) testbed and experimentally assessed in indoor radio-propagation conditions. In addition, the performance of the proposed protocol has been compared with other C-MAC schemes, also implemented over the SDR-based framework. Performance results in terms of free-collision channel occupation rate and transmission throughput show that the proposed architecture outperforms other C-MAC protocols in scenarios with low and medium density of cognitive users per available channel. In contrast, experimental results have shown that partially observable Markov decision process (POMDP)-based solutions are more suitable when the available channels are fewer than the cognitive links.