Periodic Reporting for period 1 - 6G-GOALS (6G Goal-Oriented AI-enabled Learning and Semantic Communication Networks)
Reporting period: 2024-01-01 to 2025-06-30
The 6G-GOALS project comes at a pivotal moment in communications. Data traffic is soaring, driven by applications needing ultra-reliable, real-time connectivity. Autonomous vehicles must exchange data in milliseconds, collaborative robots require constant coordination, immersive extended reality demands seamless interaction, and mission-critical IoT relies on precise, timely exchanges. Current networks are content-blind, transmitting all data equally, even when redundant. This wastes energy and spectrum, strains infrastructures, and undermines sustainable, human-centric digitalisation. The challenge is not sending “more data faster,” but communicating smarter, transmitting only what matters.
6G-GOALS addresses this by shifting from throughput- and latency-based systems to semantic- and goal-oriented communication. Information is treated as meaningful content tied to outcomes. For humans, this means context-aware messages like augmented-reality guidance; for machines, task-relevant features such as compressed semantic maps. By embedding AI, machine learning, and semantic information theory into network design, the project enables communication that is intelligent, efficient, and task driven. It develops semantic representation and compression, time- and relevance-aware protocols, and orchestration that jointly optimises communication, computation, and energy. These innovations converge into a network of intelligence, where devices exchange only contextual, timely, sustainable information.
The impacts are significant. Scientifically, 6G-GOALS pioneers new semantic information models and metrics such as value of information, semantic fidelity, and task success rate. Economically, it enables efficient applications and strengthens Europe’s role in 6G standardisation through 3GPP, ETSI, and O-RAN. Societally, it supports sustainability by reducing data flows, energy use, and electromagnetic exposure, while enabling services in healthcare, mobility, and industry. In sum, 6G-GOALS redefines communication for the next decade. By replacing data-heavy models with semantic, goal-driven systems, it paves the way for connectivity that is intelligent, sustainable, and impactful.
6G-GOALS addresses this by shifting from throughput- and latency-based systems to semantic- and goal-oriented communication. Information is treated as meaningful content tied to outcomes. For humans, this means context-aware messages like augmented-reality guidance; for machines, task-relevant features such as compressed semantic maps. By embedding AI, machine learning, and semantic information theory into network design, the project enables communication that is intelligent, efficient, and task driven. It develops semantic representation and compression, time- and relevance-aware protocols, and orchestration that jointly optimises communication, computation, and energy. These innovations converge into a network of intelligence, where devices exchange only contextual, timely, sustainable information.
The impacts are significant. Scientifically, 6G-GOALS pioneers new semantic information models and metrics such as value of information, semantic fidelity, and task success rate. Economically, it enables efficient applications and strengthens Europe’s role in 6G standardisation through 3GPP, ETSI, and O-RAN. Societally, it supports sustainability by reducing data flows, energy use, and electromagnetic exposure, while enabling services in healthcare, mobility, and industry. In sum, 6G-GOALS redefines communication for the next decade. By replacing data-heavy models with semantic, goal-driven systems, it paves the way for connectivity that is intelligent, sustainable, and impactful.
During its first reporting period, 6G-GOALS established the foundations for AI-native, semantic- and goal-oriented networks, moving from theory and architecture to algorithms, protocols, and prototypes.
A major milestone was the design of a novel 6G architecture embedding semantic principles. WP2 defined key components, notably the Semantic Radio Intelligent Controller (S-RIC) and Semantic Engine (SE), enabling semantic extraction and interpretation across layers and introducing new metrics—semantic fidelity, value of information, and task efficiency—that move beyond throughput and latency.
WP3 advanced semantic representation and compression, extending rate-distortion theory into a rate–distortion–perception trade-off. Prototypes using joint source-channel coding and deep learning achieved semantic compression of visual and control data, cutting data rates while maintaining task performance.
WP4 developed semantic protocols for reasoning and actuation, framing information as meaningful only if relevant and timely. It introduced adaptive alignment strategies, causal semantic representations, and timing-aware designs using network calculus and temporal AI models.
WP5 initiated orchestration frameworks that integrate communication, computation, and energy. The S-RIC was detailed as the orchestrator, enabling semantic-aware feedback loops and adaptive resource allocation. Early prototypes explored semantic spectrum management, control-plane signalling, and workflow coordination.
WP6 prepared proof-of-concept demonstrations, focusing on collaborative robotics with semantic mapping and distributed decision-making. A trial platform with software-defined radios and AI hardware was scoped, and initial experiments validated semantic-aware coding.
Together, these advances—new models, algorithms, and prototypes—mark a significant step towards sustainable, goal-driven 6G systems. In the next phase, results will be integrated into full-stack prototypes and field trials to demonstrate their benefits.
A major milestone was the design of a novel 6G architecture embedding semantic principles. WP2 defined key components, notably the Semantic Radio Intelligent Controller (S-RIC) and Semantic Engine (SE), enabling semantic extraction and interpretation across layers and introducing new metrics—semantic fidelity, value of information, and task efficiency—that move beyond throughput and latency.
WP3 advanced semantic representation and compression, extending rate-distortion theory into a rate–distortion–perception trade-off. Prototypes using joint source-channel coding and deep learning achieved semantic compression of visual and control data, cutting data rates while maintaining task performance.
WP4 developed semantic protocols for reasoning and actuation, framing information as meaningful only if relevant and timely. It introduced adaptive alignment strategies, causal semantic representations, and timing-aware designs using network calculus and temporal AI models.
WP5 initiated orchestration frameworks that integrate communication, computation, and energy. The S-RIC was detailed as the orchestrator, enabling semantic-aware feedback loops and adaptive resource allocation. Early prototypes explored semantic spectrum management, control-plane signalling, and workflow coordination.
WP6 prepared proof-of-concept demonstrations, focusing on collaborative robotics with semantic mapping and distributed decision-making. A trial platform with software-defined radios and AI hardware was scoped, and initial experiments validated semantic-aware coding.
Together, these advances—new models, algorithms, and prototypes—mark a significant step towards sustainable, goal-driven 6G systems. In the next phase, results will be integrated into full-stack prototypes and field trials to demonstrate their benefits.
The 6G-GOALS project goes beyond today’s state of the art by shifting from throughput-driven design to semantic- and goal-oriented communication, where networks are judged by how well they serve tasks rather than how many bits they deliver. This paradigm has already produced theoretical advances and experimental proofs, positioning Europe at the forefront of 6G.
The project extended information theory with the rate–distortion–perception trade-off, linking efficiency with task relevance and perceptual quality, and introduced new KPIs such as semantic fidelity, value of information, and task success rate. A dedicated semantic communication testbed confirmed the benefits: DeepJSCC-MIMO outperformed conventional codecs, achieving higher quality under real channel conditions with far lower bandwidth. At the architectural level, 6G-GOALS defined a Semantic RAN Intelligent Controller (S-RIC) and embedded semantic functions into programmable networks, showing how AI-native orchestration can prioritise by meaning rather than raw data. Real-world impact is already visible in collaborative robotics trials, where cobots equipped with semantic communication and hybrid control achieve better bandwidth use, faster task completion, and higher throughput than traditional methods. The potential is profound: semantic networking cuts energy use and electromagnetic exposure, supporting Europe’s green and digital transitions, while enabling efficient services in healthcare, mobility, and immersive media. Engagement with 3GPP, ETSI, and O-RAN ensures that these concepts influence global standards. For full uptake, large-scale trials, IPR protection, and supportive regulation will be needed.
The project extended information theory with the rate–distortion–perception trade-off, linking efficiency with task relevance and perceptual quality, and introduced new KPIs such as semantic fidelity, value of information, and task success rate. A dedicated semantic communication testbed confirmed the benefits: DeepJSCC-MIMO outperformed conventional codecs, achieving higher quality under real channel conditions with far lower bandwidth. At the architectural level, 6G-GOALS defined a Semantic RAN Intelligent Controller (S-RIC) and embedded semantic functions into programmable networks, showing how AI-native orchestration can prioritise by meaning rather than raw data. Real-world impact is already visible in collaborative robotics trials, where cobots equipped with semantic communication and hybrid control achieve better bandwidth use, faster task completion, and higher throughput than traditional methods. The potential is profound: semantic networking cuts energy use and electromagnetic exposure, supporting Europe’s green and digital transitions, while enabling efficient services in healthcare, mobility, and immersive media. Engagement with 3GPP, ETSI, and O-RAN ensures that these concepts influence global standards. For full uptake, large-scale trials, IPR protection, and supportive regulation will be needed.