5G and Human-Centred AI At the Edge of Network


5G and Human-Centred AI At the Edge of Network


Smart cities, digital workplaces, robotic, transportation, logistics and next-generation industries are just a few examples of the areas where the radical evolution of the last few years are impacting people’s lives. The CELTIC-NEXT Project AI-NET ANIARA, starting in June 2020, will investigate how Network and Service Automation at the Edge, enabled by 5G and human-centred AI, will accelerate the Digital Transformation in many different contexts. Konica Minolta Digital Services R&D is leading a cluster of three UK partners from the 23 European organisations participating in the AI-NET ANIARA consortium.


The project, funded within the framework of EUREKA, is coordinated by the Swedish Ericsson and has identified a set of ambitious objectives to research and develop technologies specific for a new and more secure edge infrastructure. Main characteristics of this edge infrastructure are: a large number of edge locations; heterogeneous hardware and site configurations; resource-constrained compute environments; a mix of base technologies for virtualization platforms and transport networks; and finally supporting critical services in customized network slices.

To manage the resulting complexity, the project will take advantage of Artificial Intelligence and Machine Learning to complement or replace conventional manual and proprietary optimisation and prediction algorithms. Methods developed for other domains will be adapted and evolved to suit the needs of the network edge. Safety, trustworthiness and transparency are key requirements to actually deploy new AI-based algorithm in critical infrastructures.


As recently highlighted by our CEO Shoei Yamana “Companies will thrive if they will be able to make a real, concerted effort to protect global environment and to deliver Human-Centred AI in society.” Indeed, at Konica Minolta we have been focusing on the human factor in Digital Manufacturing in the last few years. On one side, leveraging on the manufacturing competencies of the company, our laboratories are committed to bringing more specialised components and technologies towards specific solutions with a focus on Assistive AI-based services. On the other side, the company is investing in the Digital transformation as demonstrated by our recent collaboration with NEC to introduce local 5G to the new Takatsuki R&D Building in Osaka, or the cooperation with Deutsche Telekom pioneering 5G-capable smart glass devices. So, following this path, with the ANIARA project, we believe that “Manufacturing environments can greatly benefit from the use of 5G networks”, says Antonio Dina, Business Strategy & Alliances Director in Digital Services R&D of Konica Minolta. “This project will provide the fundamental building blocks for immersive interrelated collaboration contexts”.


The digitalisation of manufacturing processes and massively interconnected machines will unleash a reliable network environment where a high volume of data is processed with minimal delay (low latency support). Even though we are still in the early developments, there is a notable business need and market opportunity to allow penetration of 5G technologies on the manufacturing floors. This is highly significant as, in order to allow mobile robots and a massive set of wireless sensors to be integrated on the manufacturing space, we must first ensure that the underlying required wireless connectivity and edge intelligence that underpin the operation of those new technologies is available. Therefore, the Distributed Cloud Intelligence (DCI) Platform as a Service (PaaS), developed by Konica Minolta, is proposed to form an optimal Information Technology (IT) infrastructure for a sustainable future that simplifies the adoption of on-prem robotics in the manufacturing settings. Being a part of the AI-NET ANIARA consortium, we will be developing AI edge cloud-based techniques embedded in the 5G infrastructure that will allow dynamic provisioning of network resources to enable efficient operation of industrial mobile robots and sensors. Through its use of cutting-edge telecommunication technologies that include, amongst others, network orchestration and management techniques, the solutions will allow the agents to be mobile, and react in an almost real-time manner.


The UK consortium is complemented by partners at the European level who will contribute to the development of network capabilities including Ultra Reliable Low Latency Communication (URLLC) essential for manufacturing 5G (Ericsson, RI.SE and VTT), AI for network slicing and orchestration (Chalmers, RI.SE and Kaltio), and manufacturing industry requirements (Opel-Vauxhall, VTT and Arctos Labs). The EU-level project, AI-NET ANIARA, is part of the flagship project, AI-NET.

The Centre for Telecommunications Research (CTR) from King’s College London (KCL) hosts leading 5G capabilities as well as large-scale programmable Internet testbed. The testbed is used for robotics and remote-control applications. KCL will bring expertise in software-defined networking, and mobile edge computing, as well as ultra-low latency communication.

HAL Robotics (HAL) is a robot software company. Their core software solution is built as an extensible framework for real-time interactions, sensor integration, and machine interconnectivity. In the project, HAL will bring knowledge of parametric design, geometry driven machine control and industry-standard communication protocols, drive interface decisions and iterate over solutions to program robots, connect them to sensors and develop appropriate APIs and communication protocols.


The focus of the UK cluster of the AI-NET ANIARA project is to develop 5G edge and distributed solutions for intelligent control, monitoring and performance enhancement of industrial manufacturing assets, process flows and in-factory product optimisation. Together with Konica Minolta, Hal Robotics and King’s College London propose a solution to enable ultra-reliable, low latency communications (URLLC) for large numbers of devices suitable for deployment of 5G in manufacturing scenarios. In this collaboration we will be driven by the needs of Robotics and manufacturing. Robots are a vital part of manufacturing, logistics and healthcare industries. The future of robots requires a software architecture that enables the interfacing of robotic systems and sensor networks together with digital platforms. Ultra-low latency and the error rates 5G is capable of aligning wireless protocols with current industrial Fieldbus standards enabling robust wireless integration. This enables the addition of sensors and other devices, e.g. edge compute, to robot cells with extremely reduced downtime increasing flexibility and utilisation rates of equipment as well as permitting increased interaction with unknown physical entities and environments. Adding robot cells to a network allows centralised programming control which in turn enables distributed manufacturing whereby a central repository of tasks is created and continually updated as those tasks are distributed between connected cells with the appropriate capabilities, capacity, geographic location etc.


Photo by ThisisEngineering RAEng on Unsplash.