What a next-gen factory looks like with smart 5G
Globally the manufacturing industry is becoming more intelligent, flexible, service-oriented, and high-end. The demand for high-performance wireless networks in smart manufacturing is on the rise, with 5G able to guarantee high-quality and flexible networking capabilities for production systems and integrate data from all links in the chain.
5G’s capabilities in speed, bandwidth, latency, reliability, massive connectivity, coverage, and security can meet the strict requirements of industrial services and promote the rapid rollout of the Industrial Internet. Forecasts show that the 5G smart manufacturing market will be worth US$232 billion by 2025 and that China will spend more than 520 billion yuan (US$74.3 billion) on 5G equipment for industry application by 2030.
Recognizing the crucial symbiosis of 5G and industry, China's Ministry of Industry and Information Technology (MIIT) announced the 5G Industrial Internet 512 Project in November 2019, which aims to boost the use of 5G in industry scenarios and in turn boost the development of 5G technologies – industry is the main use for 5G and in turn the main driver of 5G technological development.
The smart factory is an important test site for the real-world application of 5G. Fixed-line still dominates the Industrial Internet, because manufacturers previously relied on wired technology to connect production facilities. However, with 5G, wireless technology is not only more suitable for complex manufacturing environments, it’s also far less expensive than line upgrades.
That’s why short-range wireless communication technologies like Wi-Fi and Bluetooth have been widely adopted in the manufacturing industry in the past few years. However, they’re not suitable for scenarios that are sensitive to packet loss, data silos, and security risks.
5G is the cornerstone for the networkization, digitalization, and smartification of manufacturing and connected factories, as well as their diverse requirements.
The 4 key elements of 5G
For 5G network capabilities to meet the demands of smart factories, we need to consider four key elements: connectivity, computing, security, and simplicity.
Reliable connectivity is the foundation of wide access, large bandwidth, and low latency.
Powerful heterogeneous edge computing platforms are required to handle the wide variety of application scenarios in enterprise campuses.
Data security and privacy via a closed-loop data environment in enterprise campuses is a strict requirement of every enterprise.
Simplicity is an obvious preferenceof both operators and enterprises when it comes to O&M and edge computing on private networks.
Why MEC matters
Multi-access edge computing (MEC) is one of the key technologies that can support low latency and increase bandwidth rates for 5G. MEC enables applications, content, and some service processing and resource-scheduling functions of the core network to be deployed on the network edge, closer to the access point. Services can be processed at the local network edge, which boosts QoE through application, content, and network orchestration.
MEC can be broadly applied in scenarios like industrial manufacturing, autonomous driving, and smart grids, as well as in typical campuses such as factories, cargo ports, and power substations.
Creating a world first
Haier – a pioneer in smart home appliances and the Industrial Internet – teamed up with China Mobile and Huawei to create the world's first smart 5G connected factory. The partners also succeeded in redefining enterprise organizational methods, business models, and ICT deployment with AI and 5G.
Located in Qingdao, Haier's smart 5G connected factory for refrigerators creates a complex, multi-scenario system that incorporates smart devices, machine vision, AR, remote control, and cloud automated guided vehicles (AGVs) – all powered by 5G. The network boasts key technologies such as cloud native ultra-distributed architecture, dynamic intelligent network slicing, high-performance heterogeneous MEC, and a 5G virtual industrial private network. It provides key service capabilities like robust data security, simplified deployment and operations, algorithm app security, and a complete maintenance model.
The network enables flexible and efficient mass-customized production and provides managers with real-time, high-precision insights into production processes and operating status. It also automates data perception in all processes, decision-making for all events, and the iteration of full-cycle scenarios.
The machine vision system, which comprises industrial cameras, has high network requirements: upstream bandwidth of between 50 Mbps and 200 Mbps, end-to-end communication latency of less than 10 ms, and reliability requirements above 99.9999 percent.
In traditional models, the execution and processing ends of machine vision systems are local to the factory and typically use cables. This creates a large spatial footprint, high per camera cost of detection, data silos, complex line maintenance and software upgrades, and long deployment and debugging times.
Haier, China Mobile, and Huawei deployed 5G base stations and 5G MEC edge nodes that, due to large bandwidth, low latency, high computing power, and cloud collaboration capabilities, could power a centrally deployed machine vision solution provided by HCvision.
Haier can now realize cloud management, the self-optimization of algorithms, and security guarantees that ensure enterprise data does not leave the campus. It has also smashed bottlenecks in traditional machine vision such as high costs, efficiency limits, unstable quality, and complex maintenance. Due to lower costs and simple maintenance, Haier can enable fine-grained functions like detecting cracks on fridge doors and optical character recognition (OCR).
The new system has also enhanced product quality, reduced equipment downtime and failures, boosted production efficiency, and reduced investment costs.
Massive data volumes can be collected at high speed through the 5G network, with data from multiple cameras and plants quickly aggregated on the 5G MEC edge node. The resulting data goldmine can be used for deep learning and self-optimization in the cloud, greatly improving the accuracy of product detection.
The factory is the first example of a cloud machine vision system integrated with 5G and edge computing in an industrial manufacturing environment.
A range of other typical factory services have also been developed, including AR-assisted maintenance, safety behavior recognition, and AGV smart navigation.The AR-assisted maintenance service, for example, is deployed on the edge-side using edge computing, with the data computed locally. Remote specialists can assist and guide the maintenance work of on-site technicians in real-time using lightweight AR glasses, saving time and cost, and solving issues with the old system like unstable Wi-Fi signals, dizziness, and cumbersome AR glasses.
For the safety behavior recognition service on the production environment, detection software is deployed on the edge side to guarantee HD video backhaul for accurate identification. The solution solves problems like high staff turnover, complex management, and replacing missing operators in specific roles. It also reduces the complexity of producing direct-current electrical goods and the difficulty of routing cables in some old factories.
The AGV smart navigation service migrates modules that require complex computing capabilities, like positioning, navigation, image recognition, and environmental perception of AGV operations, to the 5G MEC edge node. AGV smart navigation and real-time remote control functions powered by continuous 5G coverage have drastically improved the efficiency and operating accuracy of inspections, detection, and automated recognition.
AI + 5G = Smart IoT
In the future, AI combined with 5G will result in more scenario-based IoT smart products, smart integrated solutions, and AI applications, thanks to data-based simulation modeling verification, efficient production collaboration, and accurate quality control. This will optimize productivity indicators such as production efficiency, manufacturing costs, and non-warehousing rates. It will also strengthen the competitiveness of smart manufacturing and provide replicable cases of how to evolve smart connected factories. Haier’s smart factory has already attracted much attention, both in industry and other sectors, as a model for transformation. It’s the only case from China selected by the GSMA as an edge computing PoC project for "enabling operators to deliver value in new areas beyond connectivity."
At HUAWEI CONNECT 2019, Haier, HCvision, China Mobile, and Huawei announced the world's first 5G + Machine Vision solution for smart factories. Working together, the smart manufacturer, application provider, operator, and equipment vendor have developed the first cloud machine vision system that integrates 5G and edge computing for industrial manufacturing environments. The solution provides an important reference for rolling out 5G MEC applications in other scenarios, including smart logistics, smart campus, and smart homes. I’s also an important solution for empowering cross-domain, cross-platform, and cross-industry development.
The MIIT awarded the partners’ 5G MEC Device-Edge-Cloud solution for smart factories top prize at the 2nd Blossom Cup 5G Application Contest. The project has also won other awards, including the MEC Best Use Case Award from Communications World's 2019 ICT Year-End List.
Integrated development of 5G Industrial Internet will bring huge socioeconomic benefits for the industry, and promote the digitalization, networkization, and smartification of industrial enterprises. Haier, China Mobile, and Huawei will continue collaborating on the integration of next-gen ICT like 5G, cloud, big data, AI, and smart manufacturing, and making advances in the Industrial Internet that are replicable globally.