Results
- Home
- Overview
- Results
Public Deliverables
D1.3: Data Management Plan (final)
D1.4: Regulatory, societal, ethical and gender issues report
D2.1: SoA on data spaces & secure information sharing
D2.2: User Requirements and Use Cases
D2.3: System Technical Specifications and PLIADES Framework Architecture
D2.4: Security and privacy issues related to data spaces
D2.5: Human factors across data life cycles in dataspaces
D3.3: Sustainable data generation and refinement for mobility, industrial, energy, and healthcare
D3.4: Mechanisms and techniques for injecting human knowledge into the data creation process
D3.5: Extending data re-use capacity through AI-enabled data elaboration methods
D3.7: Cross-domain data Governance ensuring Quality and Legal adherence
D4.1: Multi-dataspace integration techniques for efficient data management and re-use
D4.3: Augmenting mobility intelligence through Privacy-preserving data sharing and analysis techniques
D4.7: GDPR Compliance and Enhanced Privacy for Mapping Personal Data Flows
D5.1: Unified Abstraction Framework for Enhancing Interoperability across Multiple Data Spaces
D5.2: Seamless Cross-Domain Data Exchange for Enhanced Data Interoperability
D5.5: Empowering Strategies for Inclusive AI Model Development and Sustainable Data Maintenance
D5.6: Robust AI-Enhanced Data Transformation Ensuring Secure and Privacy-Preserving Data Sharing
D6.5: Cross-source and AI-Enabled data quality assessment and improvement strategies
D7.1: System demonstration, pilots Specification and pilot sites preparation plan
D8.1: Dissemination and communication plan
D8.2: Dissemination and Communication Activities Report (v1)
D8.3: Dissemination and Communication Activities Report (v2)
D8.4: Dissemination and Communication Activities Report (v3)
D8.5: Report on European Standardization Policy and Sustainability Landscape Analysis
D8.8: Report on European Interoperability Framework Contributions
D8.9: Report on European Interoperability Framework Contributions (final)
Publications
Abstract: The existence of large-scale datasets for various autonomous driving tasks has created an increasing need for more automated annotation processes. Especially for safety critical tasks related to vehicle-pedestrian interaction, detailed and time-consuming human-made annotation is required, in order to assure accurate perception throughout any type of operating environment and for challenging conditions. In this paper, we present an automated method for the annotation of actions of humans crossing or not crossing the road. Firstly, we utilize a highly-accurate 3D multi-object tracking pipeline that combines RGB images and LiDAR data to extract the velocity and direction of movement of each pedestrian in the surrounding environment. A drivable area extraction neural network is then utilized to segment the traversable area around the vehicle. The correlation between the two above-mentioned components in the 3D space provides an accurate indication, regarding the pedestrian crossing or not-crossing the road ahead of the vehicle. Our method is validated using a custom-made multimodal dataset with an autonomous vehicle in various scenarios of a semi-structured area. The auto-generated annotations are compared directly with the human-made labels of multiple annotators and showcase the effectiveness of our method to provide an accurate indication about the human crossing the road action.
13/12/2024-15/12/2024
The accepted and registered papers will be publication in ACM Conference Proceedings?979-8-4007-1737-6, which will be indexed by EI Compendex, Scopus, etc.
Abstract: Data is a fundamental asset for organizations. Data spaces emerge as distributed structures that promote secure and reliable data sharing. The International Data Space (IDS) protocol is currently one of the main standards in the data space environment. The growing evolution of data spaces implies the emergence of challenges associated with aspects such as digital sovereignty, decentralization, veracity, security and privacy protection. Distributed Ledger Technologies (DLTs) are emerging as information structures that can provide solutions to these challenges. This paper proposes the migration of trust entities in the IDS architecture, such as the Clearing House, to Hyperledger Fabric Blockchain infrastructure as a solution mechanism to the above challenges. To this end, it presents a Hyperledger Fabric Blockchain interface that guarantees the interaction between an IDS Connector and the blockchain, which is demonstrated in this study with an Eclipse Dataspace Components (EDC) Connector.
As data becomes a valuable asset for organizations, the challenge is no longer gathering vast amounts of information but refining and managing it to generate value. To this end, there is a growing importance of transforming raw data into high-quality data products within Data Spaces, which are critical components of modern digital ecosystems. The complexity lies not only in the diversity of data sources, formats, and systems but also in the need for data products to remain adaptable and interoperable across various environments. On top of this, Data Spaces often require strict adherence to specific syntaxes and structures. In addition, poor data quality undermines trust and decision-making, and the lack of clear frameworks for processing and consuming data products within these spaces adds technical overhead. The main contribution of this manuscript is a reference architecture designed to facilitate the creation of high-quality, interoperable data products within Data Spaces. Additional contributions include an analysis of the required data types to ensure compatibility with real-world use cases, as well as addressing issues related to data quality, interoperability, and technical integration. The paper concludes with a discussion of future works and potential improvements.
Abstract: This paper presents Camera-LiDAR Fusion Transformer (CLFT) models for traffic object segmentation, which leverage the fusion of camera and LiDAR data using vision transformers. Building on the methodology of visual transformers that exploit the self-attention mechanism, we extend segmentation capabilities with additional classification options to a diverse class of objects including cyclists, traffic signs, and pedestrians across diverse weather conditions. Despite good performance, the models face challenges under adverse conditions which underscores the need for further optimization to enhance performance in darkness and rain. In summary, the CLFT models offer a compelling solution for autonomous driving perception, advancing the state-of-the-art in multimodal fusion and object segmentation, with ongoing efforts required to address existing limitations and fully harness their potential in practical deployments.
Abstact: Automated driving has advanced significantly through the use of black-box AI models, particularly in perception tasks. However, as these models have grown, concerns over the loss of explainability and interpretability have emerged, prompting a demand for creating ’glass-box’ models. Glass-box models in automated driving aim to design AI systems that are transparent, interpretable, and explainable. While such models are essential for understanding how machines operate, achieving perfect transparency in complex systems like autonomous driving may not be entirely practicable nor feasible. This paper explores arguments on both sides, suggesting a shift of the focus towards balancing interpretability and performance rather than considering them as conflicting concepts.
