*Article written by Dr Eusebiu Catana, 5G-LOGINNOV project coordinator
The 5G-LOGINNOV project exemplifies an exceptional approach to designing an innovative framework that addresses the integration and validation of CAD/CAM technologies in the domains of Industry 4.0 and ports. By doing so, it opens up new possibilities for innovation within the logistics value chain. The project’s primary focus is to support port applications by implementing key 5G technological advancements. This includes the development of cutting-edge 5G terminals specifically designed for future Connected and Automated Mobility, novel Internet of Things devices that leverage 5G capabilities, advanced data analytics, next-generation traffic management systems, and emerging 5G networks. These advancements aim to equip city ports with the capacity to effectively handle the imminent and future challenges related to traffic, efficiency, and the environment.
How it is done
Management and Network Orchestration Platform (MANO) The 5G network and cloud infrastructure are needed to be designed and deployed on the premises of any port. To support strict port security requirements, commercial Mobile Network Operator (MNO) infrastructure should be extended with Multi-access Edge Computing (MEC) capabilities that will assure smart routing of the port-related network services and applications traffic directly to the port operations support systems. In addition to commercial MNO services, the private 5G mobile network with dedicated cloud infrastructure should be built and tailored to the needs of port operations and targeted applications. 5G technologies will enable the use case innovations exploiting the eMBB service and low latency transmissions of the cellular infrastructure at the port premises, including MANO-based services and orchestration, pioneering far-edge computing services, computer vision and AI/ML video analytics. In 5GLOGINNOV a 5G edge processing node is implemented to support STS crane operations, whereas additional edge processing nodes and 4K cameras will be exploited. Massive 4K (uplink/downlink) live video transmissions towards the (far-)edge processing nodes will serve as the input of the developed ML models delivering the envisioned services. Such uplink-data-intensive applications call for enhanced capacity that cannot be served with legacy LTE networks. Hence, 5G-NSA cellular communications exploiting the eMBB service of 5G technology are needed to ensure the successful operation of the envisioned use cases. Additionally, low latency transmissions based on 5G capabilities will be exploited to deliver rapid alerts to distinct platforms/systems. Furthermore, telemetry data will be exploited (and transmitted over 5G) from several data sources onboard yard trucks (CAN-Bus, custom sensors etc.) that will be used by the predictive maintenance algorithm where 5G technology will be exploited providing a flexible, reliable and predictable environment to remotely keep track of the connected assets on a real-time basis, i.e., end-to-end monitoring of assets performance in all phases of daily port operations. Coordination with external trucks if also foreseen, dealing with external trucks inbound at the ports in order to expedite container handover operations (transition of containers from external to internal trucks and vice versa), providing an estimated time of arrivals/departures at/from the port gates etc.
5G&AI enabled rapid Alert System in Yard Truck Operations for Collision Avoidance This application is tailored to provide a rapid alert delivery system for collision avoidance between yard trucks and people. It exploits the eMBB service to transmit 4K video streams in both uplink and downlink direction, as well as the low latency capabilities of 5G to deliver rapid alerts (i.e., person detected in proximity) to the truck driver. It entails the installation of a 4K camera on the yard truck, as well as a 5G modem to establish cellular communication within the port. The camera will be oriented to potential driver’s blind spot (i.e., view angle), and transmit in real-time 4K video streams (uplink) to a GPU-enabled edge computing device. An AI-enabled service deployed in the edge processing node will receive and process the video feed to infer the presence/absence of people in truck proximity. In case of a negative event, e.g., no person detected in proximity to the truck, a blank screen is shown on a tablet installed in the truck’s cabin. For a positive event (i.e., person detected in proximity), live inference/annotated 4K video streams (downlink) are delivered to the driver in order to increase its situational awareness and avoid potential collision/accident events, delivering rapid alerts. The use case is tailored to increase personnel/people safety within the port premises, exploiting the broadband and low latency communication capabilities of 5G technology.
Optimal Surveillance Cameras and Video Analytics Frequent incidents involving boom collisions, gantry collisions or stack collisions, along with the presence of stevedoring personnel in port areas, make the risk for serious bodily injuries considerable. This use case aims at determining human presence in restricted areas (e.g. railways, areas with increased crane operations, etc.) and thus minimizing the risk for serious bodily injuries. 5G-IoT devices will be deployed at selected risk areas, equipped with a high-resolution camera (e.g. 4K, UHD), to locally perform video analytics tasks. eMBB service of 5G technology will be exploited for consuming 4K surveillance video streams. Additionally, innovative machine learning (ML) techniques will be developed and deployed at the 5G-IoT device for human presence detection. Hence, the inference accuracy and inference time of the machine learning model is of great significance for realizing the objectives of the use case. In addition to the fact that this use case increases safety measures in the employees’ workplace, it also opens up opportunities to optimize (and/or redistribute) the use of human resources in different port operations, e.g. by reducing the patrol frequency at the risk areas (currently frequent patrols are distributed to inspect risk areas), as this service is automated by the use case.
Automation for Ports: Port Control, Logistics and Remote Automation Operating port machinery (STS crane) will be equipped with industrial cameras for capturing and transfer of Ultra-High Definition (UHD) streams to the cloud-based video analytics system for identification of container markers and detection of structural damage of containers using advanced AI/ML based video processing techniques. Each targeted STC crane will have up to five cameras installed, so 5 different angled images will be received from each container. In addition, the transfer of remotely collected information will be enabled and made available to other port operations systems. Telemetry data will be collected from some of the vehicles (e.g. terminal tractors). This information will be collected from the vehicle CAN-Bus, using the IoT Device, and transmitted via the 5G network, to the port operation support system. Typical data to be collected include vehicle position, battery level, fuel level and consumption, oil level and tire pressure.
Predictive Maintenance Predictive maintenance is a significant contributor to increasing operational efficiency and reducing unplanned downtime of expensive equipment by identifying and solving problems before they occur. A key concern of almost any port is storing and managing bulky assets (such as spare/repair parts) that occupy significant space of the port, especially at ports operating close to maximum annual capacity. This use case will equip yard trucks with 5G access points connected to the truck’s data sources (CAN-Bus, GNSS, and other on-truck sensors) that will be transmitted via the 5G network to the port operations management platform. The accumulated telemetry data will be exploited by the predictive maintenance tool to potentially predict possible breakdowns, reduce downtime for repairs and optimise stock of spare parts, increase the service life of yard vehicles and optimise operational efficiency through minimisation of breakdowns. The proposed tool will capture historical and recent status data for the assets in question, utilized by the ML algorithm and driving a per yard-vehicle data-driven approach (schedule of purchases, storage of parts, proactive maintenance), by taking advantage of 5G technology that provides a flexible, reliable and predictable environment to remotely keep track of the connected assets on a real-time basis.
Mobile Core: 3GPP R15/16/17 with DSS From a core network evolution perspective, there are two main steps to supporting 5G New Radio (NR). The first step – a 5G Evolved Packet Core (EPC) with 5G NR Non-Standalone (NSA) operation – is to move forward from the existing EPC. This is the current situation implemented by the 5G-LOGINNOV Hamburg pilot site -5G production network Deutsche Telekom AG – 3GPP R15 with DSS. In the 5G NSA approach, the existing 4G core (EPC) is working as an anchor network mainly for signalling purposes. This EPC is combined with new extended radio functions – focused on the provisioning of additional mobile bandwidth capabilities (5G New Radio – 5GNR). T-Mobile / Deutsche Telekom is using additional frequencies from old UMTS solutions (2,1 GHz band) to offer more capacity for the clients. This function (dynamic frequency usage) is adapted from 3GPP R16. MEC for example in the Hamburg pilot will establish a V2X information system by combining 5G functionalities with GLOSA (Green Light Optimal Speed Advisory) to enable automated truck platooning. The optimised trajectory planning for automated vehicle manoeuvring across intersections enabled by real-time information on current and predicted traffic light signalling will require reliable connectivity and analytic capability with a low latency below 10ms. By using a MEC product by Deutsche Telekom (MobilEdgeX) between the 5G network and the connected vehicles with reducing network transfer delays to meet the specific ultrareliable and low-latency requirements necessary to serve automated truck platoons. The MEC will bring the analytics much closer to the connected vehicles by processing and combining mission-critical traffic information with manoeuvres of the vehicles and infrastructure data from the cloud. Efficient and safe driving inside a platoon requires information to be shared among the platoon as synchronously as possible. The following vehicles should be on time aware of relevant actions of the leading vehicle (imminent reduction/increment of speed), otherwise, unnecessary braking or the breaking apart of the platoon cannot be prevented.
Precise Positioning Firstly, the current applications have an accuracy of the position within an error bound of lateral of 0,57m (0,10m for 95%) and longitudinal of 1,40m (0.48m for 95%) on freeways and the conventual GNSS position information is sufficient. Secondly, the given position has to be provided in a high frequency and a low latency to be reliable in a fast-moving vehicle. Therefore, it is needed to combine uRLLC with the precise positioning service Skylark that provides accuracy for the position of up to 0.10m.
5G GLOSA & Automated Truck Platooning (ATP) The basis for proving and demonstrating the effectiveness of Green Light Optimal Speed Advisory (GLOSA) in interaction with Automated Truck Platooning (ATP) is, in addition to a safe and target-oriented communication strategy and environment, the automatic recognition and evaluation of the emission impact of driving manoeuvres, and the related influence of the infrastructure, TMS systems and TMS GLOSA measures. The driving manoeuvres are classified into characteristic cases (braking, accelerating, constant speed) and linked to the static infrastructure characteristics (curve, uphill, downhill); in parallel, the dynamic traffic control systems (traffic lights, lane and speed displays) are recorded/localised (and located as specific GLOSA POIs) and the specific information need/available content is queried and structured.
What’s next for the project?
Even if they are still at the early stage, 5G-LOGINNOV has already planned to test a number of B5G/6G candidate technologies as follows:
AI-Enabled Networks It is expected to integrate also artificial intelligence (AI) and machine learning (ML) into network operations, enabling more efficient and intelligent network management, optimization, better network security and more advanced automation of network operations. AI-ML algorithms could be used to automatically adjust network parameters in real time based on changing network conditions and user demand, as well as detect and mitigate security threats.
Massive MIMO It’s also expected to use massive MIMO (Multiple Input Multiple Output) technology, which uses a large number of antennas to transmit and receive data simultaneously. This could significantly increase network capacity and improve spectral efficiency, enabling more devices to be connected to the network simultaneously and reducing interference between devices
Dynamic Spectrum Access in the near future it’s expected to support dynamic spectrum access, which enables flexible and efficient use of available spectrum resources. Dynamic spectrum access could enable networks to dynamically allocate spectrum resources to different users and applications based on changing demand, as well as enable more efficient coexistence of different wireless technologies and services.
Network Slicing It is expected to support network slicing, which allows multiple virtual networks to be created within a single physical network infrastructure. This could enable network operators to offer more customized services to different types of users, with different performance and latency requirements
Edge Computing It is expected to support edge computing, which involves processing data and running applications at the network edge, closer to where the data is generated. This could reduce latency and enable new applications that require real-time data processing, such as autonomous vehicles.
Integrated Satellite-Terrestrial Networks It is important to support seamless integration between satellite and terrestrial networks, which could enable global coverage and connectivity for a wide range of applications, including autonomous vehicles, remote sensing, and disaster response.
Quantum Communication To use quantum key distribution (QKD) in order to encode and transmit information represents the next step to implementing security & privacy for communication. While still in the early stages of development, quantum communication has the potential to offer unprecedented levels of security and privacy for 6G communication, as well as new applications in areas such as cryptography and distributed computing.
Massive Internet of Things (IoT) connectivity To connect seamless millions of IoT devices, enabling a range of applications in smart ports, autonomous vehicles, autonomous drones, vehicle robots and precision positioning then the Massive IoT is needed.
Augmented reality (AR) and virtual reality (VR) B5G/6G could enable high-speed and low-latency communication, which could enhance the experience of AR and VR applications for seals & containers.
Energy-Efficient Communication It is needed in order to support much higher data rates and more connected devices than 5G, which could lead to increased energy consumption and carbon emissions. To address this, researchers are exploring new techniques for energy-efficient communication, such as using machine learning algorithms to optimize network operation or developing new power-efficient wireless transceivers.
Environmental monitoring B5G/6G networks could be used to monitor the environment in real-time, providing insights into climate change, air quality, and other environmental factors using vehicles as IoT sensors on the road network.
5G-LOGINNOV is the best example about how to design an innovative framework addressing integration and validation of CAD/CAM technologies related to the industry 4.0 and ports domains by creating new opportunities for LOGistics value chain INNOVation.
Kind contribution of the 5G-LOGINNOV project coordinator, Dr Eusebiu Catana.