5G & Beyond (5G+)

In this portion, we focus on developing advanced features in 5G radio access network (RAN) to support successful 5G deployment in the industrial environment such as port, ship yard, factories, construction sites, etc. For successful 5G deployment in these industrial applications to support industry 4.0 (I4.0), we need to address the following challenges:

• The Key Performance Indicator (KPI) challenge – The industrial applications have various use cases, demanding support for high data rate, dense deployment, and ultra-reliability low latency communications (URLLC) by one network, at one site.
• The Radio Wave Propagation challenge – The metallic and concrete structure in the industrial environment introduces severe blocking for radio wave propagation, creating dead zones and blind spots. The industrial activities such as movement of machines, containers objects, autonomous systems such as robots, autonomous ground vehicles (AGVs), autonomous aerial vehicles or drones lead to a highly dynamic time-varying fading environment, and electro-magnetic interferences (EMI).
• The Radio Spectrum Scarcity and Segmentation challenge: With the vision of being a key enabler of the future digital world, multi-gigahertz of bandwidth is needed, calling for more complex carrier aggregation of the licensed bands in the licensed spectrums in sub-6GHz and millimeter-wave (mmWave) bands, as well as that in the unlicensed bands.
• The Intelligence challenge – 5G is a highly sophisticated and complex system, which is not only hard to develop and deploy, but to optimally consume due to its extreme flexibility, providing countless options. Native adoption of machine learning (ML) and artificial intelligence (AI) in network management and optimization, both at the cloud server and at the edge, will be instrumental to 5G.
• The Open Architecture challenge – An open, virtualized, and fully inter-operable RAN is essential to help build a supply chain eco-system of innovative new products that will form the underpinnings of the multi-vendor, interoperable, autonomous, RAN. 
  

The goal here is to address the above-mentioned challenges by designing a Reconfigurable, Intelligent, Open, 5G RAN with the following features:

• Flexible support of carrier aggregation in both licensed and unlicensed bands under the architecture of Cloud-RAN (C-RAN) and O-RAN, with which the new radio (NR) in unlicensed bands (NR-U) is supported. This addresses the spectrum and KPI challenges. Software-defined reconfigurable radio NR-U, and carrier aggregation of licensed and unlicensed bands will bring us the advantages of increased network capacity, throughput, and user density per unit area. SDR-based NR-U will enable a localized “private” radio access network to support industrial applications.
• Embracing machine learning (ML) and artificial intelligence (AI) more natively for radio resource management and optimization, to enhance the performance toward supporting massive URLLC. This addresses the intelligence and KPI challenges.
• Scalable integration of distributed antenna network (DAN), with cloud-centric CU and baseband processing unit. This addresses the propagation and KPI challenges.
• Compliant with O-RAN Architecture, aiming to support flexible split of Control Unit, Data Unit, and Remote Unit (CU-DU-RU) functional blocks. This addresses the open architecture challenge, and supports to address the intelligence challenge.
• The SDR-based NR-U design, compliant with the C-RAN and O-RAN architecture will give us the advantages of integrating the DU, RU, and CU flexibly to overcome the radio wave propagation challenge in industrial environment, yet with intelligent radio resource and interference management, advanced cooperative processing, enabled by AI and machine learning, implemented at the CUs.
  

Future environments, be they a factory of the future or smart networks, will rely on sensors, Internet of Things (IoT) devices, mobile connectivity and applications that are designed to enhance productivity. Organizations are expected to beef up or supplement their communication infrastructure to meet the growing bandwidth requirements as they seek to manage and coordinate their systems in a holistic approach. The next generation networks will still be a combination of system covering both wired and wireless communication solutions, from private and public communication providers, offering virtualized and physical communication functions. Overall, wireless LAN usage will still be prevalent as spectrum is licensed free and usually localized within an organization’s premises. The use of 5G communications, especially private 5G networks may become prevalent in future smart environment. With the use of technologies such as network slicing, 5G networks can help in meeting the strict performance requirements needed to control thousands of fast-moving nodes (e.g. robots) from a single base station in a very dense area. With priority and pre-emption incorporated into 5G slice, higher QoS can be accorded for a particular application or service.

With heterogeneity expected and a need for a solution to exploit the diverse network resources to support the various application and services in an automatic manner, we are developing a harmonized network slicing framework covering all the current and emerging access technologies. One of the main issues to be addressed in the slicing framework is the dynamic resource allocation. Resource allocation of network slicing faces various challenges in terms of isolation, customization, elasticity, and end-to-end coordination, which are not straightforward due to the varying communication environment. We need an effective approach to convert the network service requirements to the desired network resources with considerations at different levels, including the control level, data plane level, and network wide level. To support the diverse application data needs in terms of latency, resilience, coverage, and bandwidth, an intelligent orchestration will be required to help create the network slice that will be able to support an application’s need throughout the lifecycle of the application. During run time, managing the elasticity of network slicing to guarantee the required SLA is challenging. Dynamic allocation of shared resources may affect the network performance of network slices and will require automation during operation so that application sessions can remain running with the desired QoS levels. To address the needs for intelligent orchestration and dynamic resource allocation, fast optimization algorithms, machine learning and AI techniques will be introduced into the harmonized network slicing framework.

The harmonized network slicing framework is envisioned as shown below where the structure is adapted from ETSI’s Network Function Virtualization (NFV) and management and network orchestration (MANO) framework. The scope covers:

• Architectural development catering for slicing in hybrid 5G & traditional access networks by referencing ETSI’s NFV and MANO framework.
• The development of flexible slice templates & descriptors that can be used by the various modules in the framework such as the Orchestration, Service Broker and Dynamic Slice Optimizer during runtime and facilitate arbitration.;
• Development of an automated orchestrator that delivers efficient slice creation with guarantees for isolation and delivery of optimal service to the users.
• Service Broker with Machine Learning-based slice mapping algorithms for design-time service negotiation.
• AI-based dynamic slice optimization and policies that caters for optimal real time service level adjustment.

Here, we work on leveraging 5G’s new capabilities and features to enhance the cybersecurity of Enterprise applications, for advanced manufacturing and smart urban solutions.

Our focus areas are:

1.  Zero-Trust for 5G Enterprise applications
2.  Cybersecurity for latency-aware applications.
   

Here, we aim to work on these needs/challenges:

• Innovate in user equipment (UE) terminal modem for ultra-reliable and machine type communications to tackle inaccessible RF coverage and higher path loss in FR1 (3.5GHz as compared to 4G bands).
• Develop specialized 5G wireless modem module for 5G-AMSUS platform & solutions
• Existing 5G modem chipset cannot adaptively reconfigure RF beam for sub-6GHz
  

The approach is to design a 5G modem module & platform with leading commercial chipset and customise to specific industrial applications (drone, robotics, etc), and develop an innovative ultra-wideband Signal Booster in the FR1 band to enhance link reliability of the mobile terminal using adaptive beam capabilities.

The outcomes aimed for are to demonstrate an advanced 5G UE (SA & NSA) modem solution for advanced drone surveyor grade lidar system for land survey & building construction industry. To date, our current  5G prototype solution was field tested with a leading surveyor for surveyor grade drone lidar 3D mapping at a land survey and a construction site. Our modem has been successfully tested in 5G with local MNOs sub-6. We also have validated with in a 5G teleoperated mobile vehicular platform. Our current UE can operate in 5G,4G and 3G.

The objectives here are to set up a Private 5G Testbed to demonstrate a conceptual flexible factory use case for advanced manufacturing, and to integrate the outcomes from other efforts in the programme into the test bed to demonstrate their unique differentiating features compared to the 5G offerings from commercial vendors. 

The testbed would consist of an R&D type gNodeB, Open5gs/5G core simulator and communication & computing platform to set up an integrated private network for enterprise type applications 

The testbed would demonstrate dynamic orchestration for ensuring Quality of Service in advanced manufacturing use cases in a 5G + WiFi Network with guaranteed End to End Security & Dynamic interference management with smart beam steering. We will also explore edge computing concepts to facilitate low latency applications for manufacturing use cases. The targeted use case involves the movement of parts between different work cells involving static and mobile robots.

Time Sensitive Networking (TSN) is a key emerging technology for providing a single communication network for both critical communications (with stringent performance metrics) and normal best effort communication to realize the full benefits offered by Industry 4.0. 

There are standardization groups working on developing a comprehensive set of specifications to ensure interoperability and consistent performance between the products from different vendors. TSN replaces dozens of incompatible machine-actuator-sensor-controller networks with a single integrated network, and this will provide benefits like cost / complexity reduction, interoperability, manageability, reliability, and security. When fully implemented, the network would guarantee throughput and end to end latency with specified jitter limits to ensure that critical communications can meet their needs consistently. 

Many emerging use cases of Industry 4.0 require flexibility to reconfigure manufacturing lines as well as ability to execute closed loop control involving multiple independent actors including mobile ones. Wireline TSN will not be able to provide the required flexibility. Wireless is going to be a critical element to realize this. Besides, there are other use cases where one requires accurate time alignment between widely dispersed nodes distributed over a large area. For e.g.:

• Accurately time stamped data for developing an integrated picture
• Sequential operation of independent actuators
• Collaborative Robots / AGVs / Humans
• Motion control applications like the ones using slip ring to extend wired control. Industry is keen to replace these arrangements to enhance reliability and flexibility.
  

An integrated wired + wireless TSN network with wireless extension is essential for meeting the requirements of these use cases. This can be done only with a high-availability/reliability wireless connectivity like the one provided by 5G URLLC. Even with 5G, there are many unknowns in realizing a factory grade network and this effort aims to look at the challenges and bottlenecks in realizing such a wired + wireless TSN network. A testbed would be built to evaluate the current state of the art and identify potential enhancement opportunities to take it forward.