Ultra-Reliable 5G Network for High-Level Autonomous Driving

With the proliferation of intelligent vehicles on urban roads and the emergence of automated applications such as unmanned delivery and street sweeping, autonomous driving is gradually transitioning from a mere concept to a remarkable reality. As a crucial pillar of economic development, the transportation industry is actively seeking transformation by seamlessly integrating vehicle, road, network, and cloud intelligence to empower smarter cities. Autonomous driving is no longer merely the job of individual vehicles.

High-level autonomous driving refers to a mode where machines autonomously undertake most driving tasks, truly liberating drivers’ hands, eyes, and feet while bringing unparalleled convenience into our lives. However, this also imposes higher demands on automation precision.

The advancement of vehicle intelligence is currently impeded by the lack of synergy with artificial intelligence in other domains. High-level autonomous driving vehicles require an extensive array of on-board cameras, radars, and computing chips, which can be quite costly. In terms of safety, the limited field of view and perception distance of on-board sensors may result in blind spots and delayed awareness. In addition, single-vehicle decision-making lacks efficient coordination between vehicles and road systems, leading to inaccurate predictions in complex scenarios. In consequence, single-vehicle intelligence has remained stagnant at L2 for an extended period with significant challenges for advancement.

With the advancement of 5G technology, the industry envisions harnessing the potential of this cutting-edge technology to enhance both quality and efficiency in autonomous driving solutions, propelling continuous improvements in autonomous driving through 5G+ vehicle-road-cloud integration.

ZTE has introduced the industry’s first ultra-reliable 5G network guarantee solution for autonomous driving (Fig. 1), dedicated to providing low-latency and trustworthy service connections for autonomous vehicles. The solution incorporates two major highlights: ultra-reliable 5G transmission and ultra-reliable computing service.

Ultra-Reliable 5G Transmission

5G naturally boasts numerous advantages over the conventional PC5 vehicle-road transmission system. The latter is limited to a mere 20 MHz bandwidth, posing challenges in terms of future evolution and commercialization capability. In addition to 5G’s inherent large bandwidth and low latency, ZTE has also employed various network guarantee techniques to further ensure transmission reliability.

Firstly, the 5G dual-modem gateway enables dual-link transmission and selective reception, effectively mitigating potential delays caused by single-link package loss that could compromise both driving safety and efficiency in autonomous vehicles. Additionally, to ensure the independent transmission of both links, two distinct frequency points are adopted, thereby minimizing interference and maximizing the bandwidth resources of V2X services. For instance, a 100M private V2X network can be virtually created on the existing 100M public network, restricting public network users from accessing the private V2X network, thus enhancing transmission reliability. Moreover, through intelligent pre-scheduling, required service resources and periodicity can be accurately identified, allowing the network side to proactively provide dynamic resource reservations for terminals, thereby reducing service latency. Overall, ZTE’s solution achieves a low latency of 20 ms and a high reliability of 99.99%, demonstrating industry leading performance for driving safety.

Ultra-Reliable Computing Service

In traditional methods, data is transmitted through one node but processed and computed at the other location, such as a central cloud or regional computing center. This results in prolonged transmission delays and increased investment. ZTE has innovatively merged 5G network and computing capability by simply inserting a computing board, NodeEngine, into 5G base stations. This allows service data to be processed locally in real time, rather than being transmitted over long distances, thereby accelerating decision-making processes. NodeEngine boasts plug-and-play functionality, delivering exceptional performance and ample capacity to enable real-time processing and sending of roadside perception data. This deployment method eliminates the need for new machine rooms, leveraging existing base station resources to reduce the deployment costs and the deployment time. Moreover, computing power can be shared across base stations to form an edge cloud, ensuring uninterrupted service delivery. It is estimated that a single base station modification will enable all base stations at 16 crossroads within one square kilometer to provide V2X services.

Paving the Way for Future Applications

With this groundbreaking, ultra-reliable network guarantee as the cornerstone, combined with cutting edge high-precision roadside perception and a digital-twin cloud control platform, autonomous vehicles will gain access to comprehensive and reliable V2X information. Lidars, 4D imaging radars, high-definition cameras, and other sensors are equipped on the traffic light stand for real-time, beyond-line-of-sight perception. The cloud control platform is a massive data source that can provide perceptual data for autonomous driving companies to facilitate better vehicle perception, prediction and planning. It is estimated that if every crossroad in Suzhou is connected to the platform, 198 PB of data will be generated per day, equivalent to the daily amount of data collected by 290,000 autonomous vehicles using on-board sensors. Automobile companies can also purchase this data to develop and deliver their own driver assistance services, thereby enhancing their core competitiveness. Moreover, the platform can serve as an overall monitoring center that inspects conditions of vehicles and roadside equipment, enabling the government to enhance road efficiency through trend prediction and congestion optimization.

Currently, the ultra-reliable 5G network guarantee has been trialed in the pilot intelligent road area spanning 162 kilometers, where L2 vehicles can achieve L4 autonomous driving services solely through roadside sensing and reliable 5G transmission, without the need for onboard sensors. Three Robobus lines have also been commercialized on the 5G intelligent road. The self-driving bus is empowered and operated with real-time roadside perception data and is remotely monitored by the cloud control platform. There is no bus driver behind the steering wheel of the L4 Robobus, only a safety supervisor who awaits nearby in case of any unforeseen emergencies. Additionally, over 1,000 self-driving vehicles from more than 20 automobile companies are connected to the system for real time supervision, both in the technology park and on the intelligent road. These vehicles include RoboBus, RoboTaxi, automated delivery, sweeper, patrol, sightseeing shuttle, truck, and retail cart.

With more than 3.1 million 5G base stations in China, a solid foundation has been established for the future rollout and large-scale commercialization of the integrated 5G vehicle-road-cloud system. Looking ahead, the ultra-reliable network is poised to enhance the autonomous driving capabilities of diverse automated applications, ultimately bringing an improved and more intelligent urban lifestyle to people.