Towards that vision, 6G BRAINS aims to bring AI-driven multi-agent DRL to perform the resource allocation over the high dynamic ultra-dense D2D cell free network with new spectrum links including THz and OWC to achieve up to 100 devices per m³ network density, up to 99.999% reliability and up to 0.1 ms air interface latency for the future industrial network. 6G BRAINS project is the first project to propose a comprehensive cross-layer AI driven resource allocation solution to support the massive connections over D2D assisted high dynamic cell free network enabled by THz/OWC and high resolution 3D SLAM of up to 1 mm accuracy. In order to achieve this, the enabling technologies in 6G BRAINS have been divided into four major parts including the disruptive new spectrum links, the high dynamic D2D cell free network modelling, the intelligent end-to-end network architecture integrating the multi-agent DRL scheme and AI-enhanced high resolution radio-light 3D SLAM data fusion.
BOSCH has set out a vision for a B5G-enabled industrial network where every part of the production environment is fluid, except for the walls and ceilings. Industrial machines, devices, and vehicles will be made mobile by 5G or B5G and made intelligent by edge and cloud based analytics, enabling factory owners to change their production lines according to demand in a short period of time. The massive connected IoT solution is envisioned to be one of the most promising drivers for many of the emerging use cases, including industrial automation as a means to deploy reconfigurable production systems, which can be torn up and down according to real-time demand. 6G BRAINS solution will have very few fixed elements, and even these elements will be intelligent. The building will be static, but flexible. Every piece of equipment in these new plants can be wireless, and mapped as a digital twin with the high resolution radio-light 3D SLAM. AI enabled Non-Public Network (NPN) slicing for 6G BRAINS networks will liberate factories from their fixed production lines. The MEC (multi-access edge computing) of 6G BRAINS will organise retrofitted connectivity and enable analytical applications like predictive maintenance even in old machines. The multi-agent DRL framework of this project will bring real-time intelligence and orchestration capabilities to help humans handle more changeable production routines. The AI-enabled massive D2D clusters’ modelling, enhanced by the multi-band channel models for the 200 GHz and 300 GHz as well as OWC band in 6G BRAINS over the high dynamic ultra-dense D2D industrial network is essential to open the path to 6G in industrial environment for THz and optical spectrum to bring the real scenarios into the controllable platform. The innovative disruptive technologies in 6G BRAINS like the Grant-free Non-Orthogonal Multiple Access (GF-NOMA) over cell free massive MIMO with underlaid D2D clusters, optimizing content placement in edge caches aiming at low delay and traffic offloading, user-centric interfaces enabling intent-based networking, new radio waveform for THz and OWC and evolved E2E directional network slicing management and orchestration (DNS MANO) allow a higher capacity and zero-perceived latency to meet massive URLLC connectivity requirements. The mobile robots use case demands very high requirements on latency, communication service availability, and determinism. This application can involve simultaneous transmission of non-real time data, real-time streaming data (video) and highly critical, real-time control data. The latter involves very high requirements in terms of latency and communication service availability over the same link and to the same mobile robot. Enhanced coverage in indoor (from basement to roof), outdoor (plant/factory wide) and indoor/ outdoor environment is needed due to mobility of the robots. 6G BRAINS will support seamless mobility such that there is no impairment of the application in case of movements of a mobile robot within a factory or plant.
