At the outset, the project's focus has been on exploring cutting-edge use cases. Therefore, Extended Reality (XR), industrial exoskeletons, adaptive manufacturing, and smart farming were selected due to their stringent demands for time-critical communication systems. Clear requirements, and in particular adaptation opportunities, were identified regarding the selected use cases. The in-depth analysis of these use cases extended beyond the examination of Key Performance Indicators (KPIs)—like packet delay and its variations—to include Key Value Indicators (KVIs), covering socio-economic benefits ensuring the 6G development is in harmony with wider societal goals.
Based on these insights, the technical work was carried out in DETERMINISTIC6G. In advancing 6G system capabilities, an analysis of packet delay and variation revealed a significant gap in Packet Delay Variation (PDV) between current 5G/5G-Advanced systems and wired TSN bridges. To address this, DETERMINISTIC6G introduced three innovative Packet Delay Correction (PDC) mechanisms to adjust the latency of packets in future 6G networks.
From a fundamental perspective, the project devised a theoretical framework to quantify predictability-a key feature in upcoming 6G system. Based on this framework, essential system design choices were evaluated to enhance predictability. On the practical side, the project devised various approaches to latency prediction using data-driven methods and data from commercial off-the-shelf (COTS) 5G networks as well as OpenAirInterface (OAI) platforms. Furthermore, Gaussian Mixture Models (GMMs) were augmented with extreme-value theory to accurately depict both bulk and tail regions of latency distributions, which is pivotal for underpinning time-sensitive communications, while temporal learning architectures like LSTMs and transformers for latency prediction were also explored. Overall, the results highlight the potential of probabilistic latency prediction in future systems.
The project also devised and explored various time synchronization architectures focusing on redundancy in clock sources. Additionally, in the security dimensions of time synchronization potential vulnerabilities were identified, and mitigation strategies were formulated.
In advancing 6G-enabled, converged deterministic communication systems, the project focused on developing the key enablers for seamless end-to-end integration across time-critical communication technologies (6G, TSN, DetNet) and compute domain (like edge computing). The project pinpointed existing gaps within various standardization efforts for both wired and wireless dependable communications, setting the stage for potential advancements. Furthermore, a novel approach has been devised to facilitate wireless-friendly E2E scheduling, by enhancing the robustness of time-driven scheduling in wireless systems. To address the integration with edge computing, a traffic-handling framework has been proposed to ensure coordinated traffic handling between cloudified applications and communication networks.
DETERMINISTIC6G has achieved notable strides in concept validation: The simulation framework was enhanced a data-driven model of 6G DetCom nodes, based on real measurements and a latency measurement framework was implemented within an OpenAirInterface-based 5G system, facilitating granular latency analyses.
Finally, the project developed a 6G architecture capable of continuously monitoring and predicting system behavior and aligning it with application requirements, providing a comprehensive end-to-end architecture for reliable, time-critical services.
