With 380 million connected cars projected to be on the road by 2021, how will designers address the increased network strain?
By Lou Lutostanski, Vice President, Internet of Things, Avnet
There is little doubt the connected car will continue to be among the biggest producers of data and therefore, among the greatest consumers of network bandwidth. Looking at the options on the horizon, I can see a number of different possibilities that could be beneficial to designers.
Today, allocation of the radio spectrum is largely determined by bands that are licensed and regulated by the FCC. This model doesn’t allow for the most efficient or effective use of available bandwidth. A more practical approach being investigated by groups like DARPA, is to build flexibility and intelligence into radio networks. This will allow them to autonomously and dynamically route network traffic to the best available frequency at any given moment.
IoT data compression technology offers another possible solution. This technology enables systems to reduce data transmission sizes by anywhere from 30 to 90 percent (depending on the use cases).
With the promise of higher data capacity, reduced latency, and better reliability, it’s widely believed 5G will be a “game changer.” In addition to the ability of adding more bells and whistles to automotive infotainment systems, the near real-time communication capacity of 5G should be the final piece of the technology puzzle, enabling fully autonomous driving (user acceptance is another battle altogether).
A common thread among these possibilities is the increasing dependence on software-defined radio technology- the ability to dynamically reconfigure the receiver for multiple radio standards, and to switch between them via software. It’s this flexibility that will enable the adaptation to fast-emerging communications standards for autonomous driving, while provisioning for the coming technological revolution in artificial intelligence, which is poised to extend into connected cars.
By Alexander M Wyglinski, PhD, Co-chair of the IEEE 5G Community Development Working Group
Today’s road vehicles are wirelessly connected with the world around us. GPS navigation, electronic tolling, tire pressure sensors, AM/FM radio, Bluetooth entertainment systems, and numerous other wireless technologies are integral components of most cars, trucks, and buses. With the rapid emergence of self-driving vehicle technology, it’s anticipated the information generated and shared by vehicles to ensure road safety will rapidly expand to the order of terabytes—or approximately several computer hard disks of information—per hour. As a result, we need a first/last-mile wireless technology that is reliable, seamless, low latency, ubiquitous, and supports high data rate transmissions.
Fifth Generation, or 5G, communication is a wireless technology that has the potential to support high-speed communications across hundreds and even thousands of vehicles over a single stretch of roadway via a combination of cellular base stations, wireless road-side infrastructure, and millimeter wave-based vehicle-to-vehicle communications. Additionally, leveraging both cognitive radio and massive multiple input multiple output (MIMO) technologies, 5G technology should be able to connect vehicles with the network cloud along with other vehicles in a reliable manner, regardless of the road or environmental conditions.
By addressing the first and last-mile issue associated with vehicle communications via a combination of innovative wireless networking architectures and using advanced wireless technologies capable of enhancing network capacity, 5G can provide that much-needed connection between vehicles and rest of the network. This is particularly important as this need for vehicular connectivity expands to hundreds of millions of vehicles across the country over the next several years.
By Ming Zhang, Co-Founder and CEO, zGlue Corp.
In 2016, Americans spent an average of 293 hours driving each year. Keeping drivers and their passengers safe, alert, happy, and productive increasingly requires intelligence with connectivity.
Wireless technologies enabling the Internet of Things (IoT) have unique functionality, performance and safety requirements that differ depending on the application—automotive being one of the fastest-growing. All IoT devices are connected through a wide range of different wireless standards, such as 4G/LTE, LTE-Advanced, LTE-M, WLAN, GPS, ZigBee, Bluetooth, and the upcoming 5G standard. In addition, telematics and vehicle-tracking applications are expected to drive the narrowband-IoT (NB-IoT) market for automotive and transportation.
Trying to ensure these varied devices work and communicate seamlessly brings new design and production challenges. Smaller, smarter IoT devices that run faster and consume less power will be essential to easing network strain. Designers will need a way to quickly develop, test, and optimize these for automotive applications—and then be able to scale them as needed for volume production. These applications will require controllers, sensors, and radios to be seamlessly integrated in small packages that can reliably support the necessary communications standards.
While a number of semiconductor and sensor suppliers offer point solutions, integrating them quickly and seamlessly today requires long-lead-time silicon-on-chip (SoC) designs, which take years and cost millions of dollars to develop. This is a viable approach for critical safety systems, such as anti-skid braking systems, but not for a variety of infotainment, safety monitoring, gesture recognition and other smart applications emerging as “IoT accessories” for a car.
Increasingly, there is a need for products and technologies that allow seamless, low-power integration of industry-standard semiconductor chips and sensors into IoT devices at low cost and with a fast time to market.
By Jeff Shamblin, Chief Scientist, Ethertronics
A multi-faceted approach is expected to be employed to reduce network strain and provide the data capacity needed for the increasing population of connected cars over the next several years. The aggregate solution will likely consist of a combination of increasing macro network capacity, efficiency, densification, and offloading, along with exploiting advanced antenna techniques in the automobile itself.
Adoption of newer LTE Advanced features such as multi-carrier aggregation, high-order modulation, and 4×4 MIMO will be integral components to increasing total macro cellular network capacity. Onboarding of new spectrum assets from recent auctions will increase available radio resources to be re-provisioned for LTE services. Additionally, operators are expected to accelerate network densification via deployment of small cells, which can help free up network capacity to serve higher-mobility applications.
To free up capacity for vehicle applications, traffic from low-mobility devices can be off-loaded to LTE-LAA and LTE-U networks that will co-exist with current WiFi 5 GHz bands. In general, this sharing of 5 GHz spectrum assets can free up capacity macro cellular network and provide new connectivity options to supplement the macro RAN. Additionally, as automobiles gain increasing sophistication in both cellular and WiFi connectivity, this will open opportunities to schedule “background” transactions—such as diagnostics data, content sync, firmware updates, etc.—during off-peak hours and through non-cellular networks such as WiFi.
In addition to the network-level enhancements, beam steering techniques can be implemented in vehicles to improve communication link gain, which will translate to better spectral efficiency. Dynamic adjustment of antenna system gain (via use of beam and radiation pattern steering techniques) can result in 2-4 dB of improved gain, which results in improved SINR for the link. This will have a direct impact on service range, data throughput rates, and total data served by the network.
By John D’Ambrosia, chairman, Ethernet Alliance Board of Directors
During its entire existence the IEEE 802.3 Ethernet Working Group has been pushed to provide the industry with lower cost per bit, higher bandwidth Ethernet solutions serving a growing diversity of applications. Bandwidth explosions, it determined in its 2012 Ethernet Bandwidth Assessment, are fueled by three factors: 1) Growth in the number of users; 2) Increased access rates and methods; and 3) increased services. With an estimated 380 million connected cars forecasted to be on the road by 2021, the potential for the next bandwidth tsunami is clearly visible. Mobile networks, already being hit by video, will be further stressed by these new roaming “end-stations.”
At the upcoming IEEE 802 July Plenary Session a request will be made to form a new Study Group to develop 50 GbE, 200 GbE, and 400 GbE Ethernet optics for reaches beyond 10 km. This will help provide solutions for mobile and metro-area networks, but the longevity of these higher bandwidth solutions to support the potential demand posed by connected cars is unclear. Designers will have to continue to improve bandwidths at all levels for the foreseeable future.
The factors noted above that drive bandwidth provide guidance. 380 million cars are clearly a lot. Insight regarding access rates can be obtained by looking at today’s state of cellular technology and WiFi, while today’s mobile and metro-area networks and their build-out plans provide some overview as to the nature of the network. The big unknowns are the services or applications, including infotainment, autonomous driving, vehicular sensor communications.
As the networking industry builds solutions for its understanding of today’s problems, it is clear that the first step in addressing the anticipated network strain that connected cars will represent is considering the potential bandwidth explosion. As with anything, understanding the problem will be the first step in solving it.
By Bob Noseworthy, Chief Engineer, University of New Hampshire InterOperability Laboratory (UNH-IOL)
This question presumes a situation I’m not yet convinced will be a reality, principally due to lack of the necessary infrastructure buildout. In a mere four years will we have 380 million “connected cars?” What is a Connected Car? How’s it differentiated from a cell phone with wheels?
Ever more capable vehicles emerge with more ECUs, enabling semi-autonomous driving to fully piloted vehicles. But ponder the meaning of Connected Car. Perhaps it’s an “Internet” connected vehicle, with high-speed bi-directional communications. Such capability enables apps to lock doors, find your car, summon your self-driving car, or stream infotainment. Cellular technologies (4G) support this today, as with mobile phones, but cellular isn’t optimal in all cases.
Connected Cars can benefit significantly simply by receiving information from roadside infrastructure. One example is the SPaT Challenge, seeking to enable 20 intersections in all 50 states with vehicle to infrastructure (V2I), or rather “I2V,” communication capability to the vehicle. Such “Signaling Phasing and Timing” (SPaT) information can enable better awareness of signals, and fuel optimization. Dedicated Short Reach Communication (DSRC) enables V2V or V2I, to support 380 million vehicles by 2021, a substantial challenge is simply the scale of infrastructure build-out still required. As the SPaT challenge highlights, a goal of only 20 intersections per state by 2020 is the current realistic target. DSRC seeks to provide reliable low-latency communications channel to vehicles for such things as crash avoidance alerts, first responders or traffic conditions. Cars with DSRC radios built in are only now emerging.
Vehicles are moving to internal high-speed automotive Ethernet networks, enabling rapid low-latency deterministic communication within the vehicle, and through gateways to communicate between vehicles, and to infrastructure. Such gateways may be wireless (5G, DSRC, or WiFi) or a high-speed wired connection in a mechanic’s shop for updates and diagnostics.
By Mark Smith, Ph.D., Product Line Manager, Automotive Sensor Interface Product, Microsemi Corporation
Today’s automotive designers focus on using local data storage and standards based vehicle to vehicle (V2V) WiFi communication technologies to optimize the connected car experience and safety for the consumer. We estimate that the connected cars will increase the network traffic by more than double what we have today. The connected car experience includes advanced navigation, voice activated services, video and audio passenger entertainment, along with traditional email and instant messaging. While all areas of the network infrastructure will need to be increased, the bottleneck will still be the cellular connection between the car and the cell towers.
The availability of high performance solid state drives (SSD) for local storage and the ubiquity of WiFi networks in the home will mitigate the bandwidth limitations of 4G/LTE networks, providing new levels of driver safety while facilitating more efficient data rates to the network infrastructure. The SSD will locally store audio and video for passenger entertainment, uploading new content when the car is at home, or in a trusted WiFi network. We also see designers utilizing this network to facilitate better cellular transfer rates. While on the road, the cellular work load can be load-balanced by utilizing the WiFi network as a mesh network. In crowded urban networks, the WiFi V2V connection enables roads to be turned into communication networks with the upload and download links being the cell towers. If a particular cell tower is overloaded, the adjacent or cell tower further down the street can provide the link to the network.
WiFi mesh networks are currently being adopted in the home environment, and we anticipate this expanding to the new V2V WiFi networks to optimize data transfer and reduce the strain on the cellular network. The connected car promises a feature-rich environment with added consumer value and improved safety and reliability.
By Thomas Scannell, Automotive Lead – Amphenol ICC
Increased connectivity requirements of today’s connected vehicles are already pushing established automotive bandwidth capabilities to their limit. As automotive technology continues to progress in adopting more advanced internal electronics and safety features, the next step in the process is to extend their connectivity to enable vehicles to communicate, interact, and respond to their surrounding environment, as well as to other vehicles and drivers. Data communications requirements are expected to expand exponentially, and in order to support these new bandwidth requirements, design engineers will need to develop a new automotive networking protocol standard.
The expanding requirements for data transmission speeds, signal integrity, and network integration in tomorrow’s connected vehicles share a similar path taken by individual computers as they evolved to networks and then on to data centers. Advancing beyond the initial automotive protocols (such as CAN bus) that were developed as a first attempt at data integration, a modified version of Ethernet has gained popularity in the automotive market as a new data communications protocol, primarily for its ability to enable data transmission rates at unsurpassed speeds (some 100X faster than CAN bus). Automotive Ethernet may prove sufficient for the interim, but designers will have to continue adapting existing technologies and developing new technologies to keep ahead of projected bandwidth requirements. Most likely, designers will continue to borrow from proven data center networking technology due to the technological parallel of increasing connectivity between vehicle subsystems and between computers in a data center.
