5G: Milestone for Autonomous Vehicles
Increased volumes of data in our increasingly mobile and connected world will necessitate a new network that can sustain the communications infrastructure being created.
Autonomous vehicles illustrate the multiple strands of connectivity that will require the coverage of 5G. They will, in effect, be a data center on wheels, with sensor fusion providing critical functions, combined with graphics processing for 360-degree views and interpretation for location data, together with the processing of data for route planning and machine and deep learning for safety features.
The European Telecommunications Standards Institute (ETSI) definition of 5G is that it should “offer a user experience near that of fixed networks with near total 5G coverage, cater for massive deployment of Internet of Things, while still offering acceptable levels of energy consumption, equipment cost and network deployment and operation cost to ensure the service can be provided economically.” This will need scalable mobility and performance, reliability, energy efficiency, and security requirements to be defined.
The heterogeneous 5G network will be made up of many wireless technologies, Long Term Evolution (LTE), Wi-Fi, millimeterwave (mmWave), Narrowband Internet of Things (NB-IoT), and a new 5G interface. Creating an infrastructure that allows all of these to work seamlessly together is the first challenge faced by the industry.
In January, ahead of CES 2017, Intel® announced the Intel 5G Modem (Figure 1). It is, said the company, the world’s first global 5G modem, intended to accelerate the development of 5G-enabled devices for early deployment.
It integrates a baseband chip with a 5G transceiver for both sub-6GHz (3.3 to 4.2GHz) and mmWave capabilities. There is also support for 5G New Radio (NR) technology features, such as low latency frame structures, which are required for safety-critical and mission-critical applications. Other features are advanced channel coding and support for 2×2 and 4×4 Multiple Input Multiple Output (MIMO) configurations, and beamforming.
Paired with the Intel sub-GHz Radio Frequency Integrated Circuit (RFIC), it supports sub-6GHz bands for deployment in Asia and Europe. For trials and deployment in the US, Korea and Japan, it can be paired with the mature 28GHz 5G mmWave RFIC to support initial 5G spectrum worldwide.
The 5G Modem’s combination of 3GPP 5G NR technology features is expected to deliver faster connectivity and increase responsiveness. The Intel 5G Modem’s support for the 3GPP NR specification is expected to drive worldwide adoption of the 3GPP 5G standard.
Increased Data Traffic
The first 5G standards are expected to be approved in March 2018, with the first specifications completed in September 2018, ready for commercial deployment.
Aicha Evans, Corporate Vice President and General Manager for Intel’s Communication and Devices Group, said the Intel 5G Modem will support early trials and lay a foundation to accelerate the deployment of products that support the 3GPP NR specification. While 5G promises to enable billions of devices to become ‘smart’ through seamless connectivity, increased computing power, and access to data and analytics that are stored at the edge of the cloud, Evans points out the need for a bridge. “Today’s communications systems weren’t designed to accommodate the massive bandwidth required for such an evolution, or the ultra-low latency needed to allow devices or even vehicles to react to split-second events,” she said.
Vehicles will generate large amounts of data and will need to receive large amounts to navigate and react to changing conditions and environments. 5G, says Evans, will deliver the latency required and technologies such as Vehicle-to-Vehicle (V2V) connectivity for autonomous driving.
At CES 2017, Intel, Mobileye and BMW Group announced that the partnership plans a fleet of 40 autonomous vehicles, which will be on the roads later this year (Figure 2). Intel provides the computing between the vehicle and the data center. Its Intel GO™ automated driving is made up of processor and Field Programmable Gate Array (FPGA) technologies in a scalable development and compute platform that handles sensor fusion, driving policy, environment modeling, path planning and decision making. In the data center, Intel GO is supported by Intel Xeon® and Intel Xeon Phi™ processors and the Intel Nervana™ platform for Artificial Intelligence (AI).
Mobileye supplies the EyeQ®5 computer vision processor, which interprets the data from the 360-degree surround view vision sensors. Mobileye is also working with BMW to create a full model of the environment surrounding a vehicle based on input from vision, radar, and lidar sensors. Mobileye’s reinforcement learning algorithms will provide AI for safety.
Mobileye was acquired by Intel in March 2017. Its processor portfolio provides pedestrian collision warning, forward collision warning, and lane departure warning systems, as well as intelligent high beam systems, and speed limit indication, which reads roadway signs and notifies the driver if the vehicle exceeds the speed limit.
The project to test a fleet of vehicles using the joint technology and in real-world conditions is a milestone, according to Klaus Fröhlich, Member of the Board of Management of BMW AG for Development. “This is a significant step towards the introduction of the BMW iNEXT in 2021, which will be the BMW Group’s first, fully autonomous vehicle,” he said at CES 2017.
Caroline Hayes has been a journalist covering the electronics sector for more than 20 years. She has worked on several European titles, reporting on a variety of industries, including communications, broadcast and automotive.