5G Advanced促进多机器人协作-中兴通讯；中国电信

Streamlined private 5G network architecture is designed to deliver robusthigh-performance connectivity to train and supervise robots Executive Summary •The development of physical AI promises to enable robots to work in a wide range of environments,such as hotels, factories, warehouses, ports and even homes.•Private 5G networks could deliver the robust and flexible connectivity needed to train robots and latersupervise them in enterprise environments.•China Telecom, ZTE, DroidUp and AgiBot have teamed up to develop a simplified and robust private5G network that uses 5G-Advanced technologies to support ultra-high throughput, ultra-low latencyand high reliability connectivity.•Their flexible EasyOn·Robot solution, which doesn’t require a core network, is designed to ensure eachrobot has an uplink of 200 Mbps.•In a demo, the connectivity enabled DroidUp robots to operate an unmanned convenience store,delivering an automated shopping experience with seamless entry, personalised recommendations,autonomous checkout, and delivery•Aiming to establish “a significant industry presence” by 2027, ZTE says the EasyOn·Robot privatenetwork solution is targeted at a global market. For enterprise deployments, connectivity will alsobe required so that a central management systemcan monitor robots in real time, allocate tasks andeven facilitate so-called swarm behaviours, in whichmultiple robots team up to complete complex tasks. Challenge–mobile robotsneedcontinuous reliableconnectivity Connectivity can also help a robot determine itsprecise position. When first deployed in a newlocation, a robot will typically leverage its owncapabilities to build a local map, which will need tobe continually updated as the environment changes.After the robot uses the map to reach a designatedlocation, it may be required to upload images, sothat the server can confirm it is in exactly the rightlocation to perform the required task. If the robotis slightly out of position, the connectivity can beused to make real-time adjustments, enabling high-precision positioning in dynamic environments. There is growing demand for mobile robots that canbe employed across a wide range of environmentsfrom factories and farms to hotels and homes.To operate effectively in different locations, robotsneed to be trained to recognise and respond to thespecifics of each physical environment. For example,in a home they might navigate a rug or a sofa,whereas in a port they may encounter slipperysurfaces or strong winds. To develop the necessary “physical artificialintelligence (AI)”, robots need robust high-speedconnectivity that can transmit the large volumes ofdata being captured by onboard cameras and othersensors to train the underlying AI models. Solution – a dedicatedprivate 5G Advanced network Even once the training is complete, a robot may stillrequire continuous connectivity both to co-operatewith other robots and to access information andprocessing power located on servers. While somerobots may have sophisticated physical AI systemson-board, others may need to frequently query AImodels running elsewhere. The balance betweenon-board computing power, cost and battery life willvary depending on the operational scenario and thespecific needs of the industry in which the robot isoperating. ZTE believes private 5G networks are well suitedto meet robots’ connectivity requirements.Whereas wired connections restrict mobility andWi-Fi networks can be subject to interference andcongestion, private 5G networks could deliverthe robust and flexible connectivity needed totrain robots and to supervise them in enterpriseenvironments, such as factories, ports, warehousesand other industrial facilities. China Telecom, ZTE, DroidUp (a humanoid robotcompany) and AgiBot (Zhiyuan Robotics) haveteamed up to develop a simplified and robustprivate 5G network that enables robots to movefreely across various scenarios. Their EasyOn·Robotsolution uses 5G-Advanced technologies to supportultra-high throughput (up to 2 Gbps uplink and 6Gbps downlink), ultra-low latency (under 10ms), andhigh reliability (99.99%) connectivity, according toZTE, thereby enabling robots to move freely acrossdifferent scenarios while maintaining real-time, high-quality data transmission. Ifthe robot is slightly out ofposition, the connectivitycan be used to make real-timeadjustments,enablinghigh-precision positioning indynamic environments. EasyOn·Robot is designed to support sufficientcapacity to deliver uninterrupted connectivity indemanding high-density environments, such as anunmanned convenience store, where robots will needto autonomously guide, recommend, and deliverproducts efficiently. By ensuring that each robot hasan uplink of 200 Mbps, ZTE says EasyOn·Robot cancontinuously capture high-resolution operational andenvironmental data in real-time. This architecture supportsscalable, intelligent, and secureconnectivity that adapts tocomplex, dynamic applications.The platform’s modular, mu