5G Enables AI to Unleash Its Vast Potential

<!–

–>

Article By : Jing Wang, Mouser Electronics

5G development and deployment are accelerating, especially in Asia. Field trials are underway, components are coming, and testing covers the spectrum in more ways than one. What are the challenges and how is the ecosystem shaping up? Find out more in this month’s In Focus series.

---

3G led us into the age of the internet. 5G now represents an even more revolutionary transformation and a leap into the future. 5G is ushering in the era of the Internet of Everything (IoE), unleashing the power of artificial intelligence (AI) to leverage the immense amount of cloud data produced by different devices to communicate smoothly with each other. The power of AI applications will always depend on the strength of their networks. As such, the high bandwidth, low latency, and boosted signal strength of 5G will undoubtedly boost AI with even greater strength.

5G will play a huge role in promoting the constant growth of the Internet of Things (IoT). It is forecast that by 2025, more than 75.44 billion pieces of internet-connected devices will be installed on a global basis. We can look forward to and even already start to experience some of the changes that 5G will make in areas such as urban living, health care, and agriculture.

Through the combination of 5G and AI, revolutionary innovations will take place in the applications of self-driving vehicles and drones. This article paints a picture of the vast and complex 5G network systems so that readers can get a clearer understanding of what 5G truly means for the future.

What Is 5G?

5G refers to fifth-generation mobile networks. The new global wireless standard follows improvements on the 1G, 2G, 3G, and 4G networks. It uses all-new wireless infrastructure to provide peak download speeds 20 times faster than 4G. It is also expected to do away with pretty much all lag time. 5G technology will officially inaugurate the era of the IoT, where billions of machines, devices, and sensors will be interconnected at low cost.

5G has three major networking capabilities: enhanced mobile broadband (high bandwidth), bulk machine communications (massive connectivity), and high-reliability communications (low latency).

• High bandwidth: The peak download speed reaches 20Gbps, while 4G only provides 1Gbps. This meets the users’ needs for high-data speeds in security surveillance, environmental monitoring, and product inspection.
• Massive connectivity: 5G’s three major features drive the development of IoT services, creating situations for lower requirements in real-time network awareness ability but higher requirements for terminal density. This will be widely applied in public utilities, manufacturing, agriculture, transportation, and power generation and distribution.
• Low latency: 5G will shorten the delay time between base stations and terminals to only 0.5 milliseconds. This capability truly differentiates 5G from other network standards, making it versatility applicable in fields that require higher reliability, such as medicine and transportation. Examples are remote device control (remote surgery) and security surveillance (urban public safety monitoring).

The Differences Between 5G and 4G

4G’s peak bandwidth speed is 1Gbps, which can meet common usage needs such as high-definition images and videos. 4G latency performance is far inferior to 5G, which explains why 5G communication networks are needed to solve 4G’s inability to support seamless real-time gameplay.

4G uses radio waves (with frequency bands of 1,880MHz to 1,900MHz, 2,320MHz to 2,370MHz, and 2,575MHz to 2,635MHz), but 5G uses millimeter waves (3,300MHz to 3,400MHz—basically for indoor use only—3,400MHz to 3,600MHz, and 4,800MHz to 5,000MHz). This shows that different frequency bands 5G and 4G uses. Although frequency bands of 4G networks are in the range of 1.8GHz to 2.6GHz, frequency bands used by 5G are higher, with the lower end of its range being above 3GHz. This creates the requirement that all equipment which will use 5G must be reconfigured.

The Connections and Interactions of 5G and AI

5G and AI interconnection and interactivity are reflected in the five areas of IoT: cloud computing, big data, edge computing, and network slicing.

• IoT: This is a network formed by interconnecting all types of information-sensing devices, where sensors capture information, the IoT network transfers information, and the IoT platform processes information and manages connections with sensors and terminals using AI technology to display information and run applications.
• Cloud computing: This provides storage and computing of bulk data and multi-user data aggregation and sharing capabilities. Cloud computing provides the basic storage and computing capabilities for the IoT and big data and provides powerful computing capability for AI algorithms.
• Big data: This is the processing of bulk data through dedicated cloud servers to optimize workflow. Big data’s true significance lies not in its vast amount of information but the specialized processing of all this meaningful data. 5G’s greatest potential benefit will be found in the generation of new data, where all sorts of devices are connected to the related networks. Big data, which could be called the core component in AI, is driving the development of AI’s new era. 5G will provide the infrastructure and massive data needed for AI’s success.
• Low latency: 5G will shorten the delay between base stations and terminals to only 0.5 milliseconds. This creates the feature of low latency. On the one hand, 5G combined with mobile-edge computing can effectively decrease data throughput by only transmitting the data and information that has been fully processed through the network. This provides AI application users with localized fast processing and analysis, which can match user demands and reduce access latency. On the other hand, the advantages of 5G can allow for the broader collection of smart terminal data, and information analysis and processing can be done using AI technology on terminals and the edge. The data is aggregated on the corresponding data platforms to allow for rapid feedback after computing and processing. Overall, edge computing will move cloud storage and cloud computing capacity toward the edge, which is closer to the user, so that edge computing and cloud computing can work hand-in-hand. The edge will store and process real-time data, and the cloud will store and process shared bulk data, thus preventing the bandwidth waste and delay caused by transmitting large amounts of data to the core network for processing.
• Network slicing: Network slicing refers to cutting up a shared physical network into multiple virtual end-to-end networks, with each slice able to possess its independent network resources while establishing total isolation between the different slices. Therefore, if a failure occurs to one individual slice, it will not affect the other slices. In a 5G application scenario, the operator can define different slices based on different service types to meet users’ different needs in delay, throughput, and capacity. The 5G network slicing technology is a fully virtual cloud architecture, with no need for building a separate network for each server, which will greatly save on costs.

What Changes Will 5G Bring to AI?

Examples discussed in this section illustrate what kind of changes are expected to occur in the field of AI because of the three main network advantages of 5G.

5G will boost AI’s the ability to comprehend language and context: The high-speed data transmission and multi-device data interconnectivity of 5G will allow AI models to maintain continuous learning and dynamic modes. Take, for example, the following specific scenario:

The smart-voice call software being developed by a certain innovative company provides real-time display of the conversation record on the customer’s screen during the call. AI can then perform emotion analysis and provide answers to the customer’s questions. Many of its functions can only be used on desktop computers at the moment. As soon as the introduction of 5G makes 5G-grade connectivity a possibility for all users, this company will start to make these functions available on mobile devices. By using 5G technology, mobile terminals will instantly be equipped with all the functions of computers connected to ultra-high-speed networks.

In terms of 5G and the smart home, different types of AI-based voice-controlled devices will be placed inside the home. These include built-in smart switches, smart voice-controlled thermostats, and humidity monitors, smart speakers, smart alarm clocks, and smart TVs. More and more terminal devices in the home will be equipped with remote AI voice control and home networking functions. The 5G smart home will adapt to different lifestyle scenarios through voice-control functions that cover any scenario and coordinated voice-control functions offering coverage to the whole house, thereby meeting the constantly growing needs of future smart living.

5G’s Effect on the Digital City

The arrival of the 5G era will allow for the simultaneous network access of all smart devices in the smart city and their voluminous amounts of data (Figure 1). It will support the virtual-real interaction of devices and the possibility of millisecond-fast responses. This will promote the firm establishment of multi-scenario AI-based services.

What V2X needs to transmit and share is not the final decisions or commands but just information about the state of the surrounding vehicles and environment. The self-driving systems of the vehicles themselves can make their own driving decisions through models generated using cloud computing and AI algorithms.

Note that the information required to be transmitted by V2X includes sensor data and the driving status of surrounding vehicles.

The Development of Remote Medical Care

Similarly benefiting from 5G’s low latency and high bandwidth, an experimental 5G remote surgery–in which liver lobe resection was performed on a piglet–achieved success. The procedure was jointly undertaken by Huawei, China Unicom Fujian, and Beijing’s 301 Hospital. The patient and doctor in this surgery were 50km apart, and 5G-based remote-controlled robotics was used. All aspects of the collaborative operation coordinated together seamlessly, with a latency of only about 0.1 second.

The image below illustrates the process of the application of 5G in a remote surgery (Figure 2). The doctor needs to obtain results from the data analysis during the surgery, and a large amount of data is stored in cloud resources for the doctor to reference. The newly generated data from the surgery in progress can be sent to cloud resources. This is sent for analysis to edge locations using the 5G core network and network slicing technology. Results are returned to the doctor after the edge computing center analyzes the data and makes predictions using AI and machine-learning technology.

Applications in Smart Farming

To put it simply, 5G’s role in farming is providing large amounts of data in agricultural science and combines it with existing farming scenarios. This then offers analyses and predictions and enables agricultural production monitoring through computer platforms.

Because of 5G’s high speed and low latency, agricultural producers can build an entire monitoring and control system that can result in precise, scientific, and efficient management of the crops’ health and growth. This is achieved through the interconnection of IoT devices that provide real-time data collection and AI models that provide rapid decision-making. For example, IoT devices can be used to collect data in the areas of atmosphere, soil, crops, and pests to guide agricultural production at any time and any place. Smart farming will use IoT technology so that interconnection and intercommunication can be realized between all farming equipment used.

Conclusion

5G technology will drive the rapid development of AI, thereby setting the stage for the Internet of Everything and an era of unprecedented data sharing. When machines can communicate with each other without human control smoothly, true data sharing, real-time data analysis/prediction, and solution analysis will happen. For example, self-driving automobiles will make decisions about changing routes based on instant predictions provided by mounted cameras that capture road conditions in real-time. The potential areas for using 5G are unlimited, as the technology will surely find its way into every industry and even every corner of our daily lives. It will contribute toward a future in which all fields are interconnected and mutually complementary, moving toward constructing a smart society. 5G now represents an even more revolutionary transformation and a leap into the future.

About the Author

Wang Jing is a machine-learning algorithm engineer currently working in the field of automotive inspection. Passionate about creating technical articles, she hopes her writings will arouse readers’ interest in artificial intelligence and inspire more professionals to combine AI with cloud technology and big data to make life safe and convenient.

---

Related