The contemporary automobile is thoroughly digital—the ultimate "mobile device."
- It hosts a variety of digital services for navigation, infotainment and internet access.
- Sensors on all its core components—engine, transmission, steering—continuously generate data to optimize performance and economy.
- Environmental sensors are added to the mix to deliver such safety features as automatic braking and lane divergence control.
- When those sensors participate in the Internet of Things (IoT), their aggregate data serves manufacturers, insurers and the driving public through smart traffic control and hazard avoidance.
The goal of fully autonomous (self-driving) cars—smart, safe and efficient—is on the near horizon.
Many of the car's digital applications depend on not only onboard technology but also external connections— for everyday services like navigation, updates to digital components and sharing data as authorized by the owner.
Keeping cars connected is a significant technological challenge. The volume of data is large and growing, as is the population of connected vehicles. Some data transmissions—vehicle updates, for example—must be verified, some require low latency (i.e., minimal delay), and all must be kept secure. Highly advanced and distributed computing and communications technologies, incorporating cloud services and edge computing, are meeting the challenge, helping to put the latest "mobile devices" on the road.
Perhaps no other industry has embraced connectivity and IoT as fully as has the automotive industry, incorporating an increasing level of functionality, flexibility and connectivity into its product offerings. In fact, some organizations are redefining themselves as "mobility" companies, offering such things as bikes and scooters with built-in locks that can be opened by a smartphone app, and with GPS, so they can be tracked and retrieved. Some of these companies have plans for autonomous delivery and other services that probably would have seemed unfeasible a decade ago.
AT&T is actively involved in providing LTE communication services to many automotive and other transportation companies, and it is also working with the automotive industry to develop standards for connected and autonomous vehicles. As of Q3 2018, nearly 24 million vehicles have been connected to AT&T's networks and the networks of its global partners. AT&T currently provides services to 29 global vehicle brands.
The connection to the vehicle (we include cars, trucks, RVs and commercial fleets) delivers value in many areas—to the vehicle user, to vehicle manufacturers, to infrastructure and, in some cases, to insurance companies and regulatory agencies. We'll look at several of them.
Infotainment and Navigation
Infotainment offerings now include more than the ability to play the radio, CDs or flash media. Bluetooth compatibility enables the music system to play audio from a user's mobile device. But it goes far beyond this. Through the cellular link, the car can act as a car phone, and some vehicles become Wi-Fi hotspots. Passengers can use the in-car Wi-Fi to provide a stable internet connection for entertainment or communication.
Navigation systems (GPS) can receive over the air (OTA) updates of maps and current traffic conditions. Future capabilities will enable vehicles to report differences between the computer map and actual road conditions, based on data captured by onboard cameras and other sensors. The data will be sent into the cloud for processing, and updated maps and route information can be sent to other vehicles in the area.
Adaptive Driver Assistance Systems (ADAS)
ADAS has been rapidly improving. Early examples include automatic parallel parking and lane divergence detection. Vehicles now use vision or lidar (laser-based detection systems) to sense stopped vehicles and apply brakes in order to avoid accidents. ADAS will utilize the megabytes of data produced by sensors that are on—and in—vehicles to provide better control of the vehicle. The ultimate goal—potentially only a few years away—is the autonomous vehicle.
"When we hear fantastic claims about autonomous vehicles being out in a few years—there are technical challenges we have to work out, and we're a ways away," says Evan Hirsh, principal of PwC strategy consulting firm Strategy&. "We don't know whether our sensors are good enough to detect what we need to know, do we have adequate processing, what communications capabilities are taking the data off the vehicle," Hirsh continues.
Vehicle to Infrastructure (VtoI) and Infrastructure to Vehicle (ItoV)
These are broad categories: These transmissions are used in many ways to benefit the user, the manufacturer and, with the owner's permission, others.
For example, the vehicle manufacturer may use aggregate data from vehicles, sent through the cloud to a central server, in order to "understand how well the car is performing and give insights into developing the next car," says Hirsh. This technology can also be used to detect or predict potential vehicle problems.
"The manufacturer wants to understand how cars are used by users, to better model and proactively predict when faults occur and develop even better diagnostic data," says John Summers, VP and CTO at Akamai Technologies.
At the Consumer Electronics Show in 2017, AT&T demonstrated how data can be used to ascertain and act on road conditions. "We showed how we can take anonymized data from certain events, such as tire slippage in a connected car, and analyze it," says Cameron Coursey, VP of product development at AT&T's IoT group. "If we know the location and there are other reports of slippage (from other vehicles in the same location), we can ascertain that there is ice on the road. We can then make this information available to drivers a few miles away, so they can plan alternate routes."
With the owner's permission, information about driving habits—where the driver goes, what times of day the driver is using the car, number of miles traveled, and specific driving behaviors such as speed and hard braking can be shared with insurance companies for rate calculation. "We give our customers the opportunity to self-monitor driving behaviors in order to improve fuel economy and become safer drivers," says Steve London, head of Connected Vehicle Solutions at General Motors. "Our customers can also privately check for discounts with insurance carriers."
Of fundamental importance to the consumer is the ability of the manufacturer to perform automatic software and firmware upgrades to vehicle systems. Tesla has been able to upgrade its vehicle software simply by transmitting the new code to its cars. But Tesla is not alone in having this capability. Other manufacturers, including Fiat Chrysler and others, have used this capability to resolve warranty recall issues or fix security issues as they're detected and resolved.
Infrastructure to Vehicle downloads are often performed using over the air (OTA) systems. Akamai, for example, has an OTA solution (OTA Updates) for automotive manufacturers that is designed to download the software updates, error-free, and to confirm that they were successfully transmitted. OTA software update campaigns ensure that all selected vehicles do get the secured software update in a timely manner.
Some of GM's applications aren't particularly time sensitive—according to London, responses for remote commands in the myChevrolet, myBuick, myGMC or myCadillac mobile app can take up to 30 seconds, depending on network speeds. In-vehicle apps leverage the 4G connectivity in the vehicle to stream content.
A Matter of Time...
With the evolution of the connected car, issues of latency have become important. The amount of time it takes for the vehicle to communicate with the infrastructure can be anything from a casual lag to a life-and-death issue—for instance, when notifying emergency services of an accident.
For example, in an infotainment system, where data from an entertainment source is transmitted at relatively low data rates, a short interruption shouldn't be much of an issue. GM, working with Amazon Prime, has an option that enables the delivery driver to get GM to open the vehicle so that packages can be put into the trunk or passenger compartment. Making the Amazon Prime delivery driver wait for a few seconds while the trunk is opened isn't a big deal.
On the other hand, in the event of a vehicle accident, if systems are configured to support it, data may go to emergency services or transportation departments. The information may go from vehicle to infrastructure, where the location is noted (from GPS data) and warnings are sent to other vehicles in the area to avoid the scene of the accident. Information may possibly be sent to the vehicle owner's insurance company, if the owner has set up this capability.
As the distance from vehicle to infrastructure increases, so does the time it takes for the signals to be passed between them. A signal sent from a vehicle to an infrastructure component just a few hundred feet away will naturally be received much sooner (in terms of milliseconds) than a signal sent to a device thousands of feet away. Further, if the device receiving the signal can process and respond to it at the device, the latency between signal and response will be at a minimum.
What we're describing here is an "edge" device. "Data from edge devices is much closer to the vehicle, the data does not have to go back to a data center to be analyzed," says AT&T's Coursey.
Akamai Technologies is a pioneer in edge computing, with more than 200,000 servers. One of its product offerings, IoT Edge Connect, collects and securely transports prescribed data to and from the connected vehicles to the requestor for further processing and analysis. "We view ourselves as enabling secure and reliable communications between the auto manufacturer and a set of connected cars," says Akamai's Summers. Additionally, communications are encrypted, and the devices are able to perform data aggregation and processing, connect to other closely spaced edge devices and provide rapid connections to the cloud.
Communications standards for connections between vehicles and infrastructure are still in a state of flux. "Two types of architecture are in debate," says Strategy&'s Evan Hirsh. "DSRC [dedicated shortrange communications] and 5G [fifth generation]. DSRC is happening right now—but given regulation and adoption [hurdles], it is unclear how long it will take to penetrate in a meaningful way."
Privacy and Security
Ever since the first reports of a hacker's ability to remotely control the accelerator pedal on a car, there's been concern about the safety of connected cars. In worst-case scenarios, hackers may even be able to attack large groups of cars or to harvest data from thousands of customers.
But these are truly worst cases, and the automakers are hard at work to ensure that security breaches never happen. Security is an "important issue that is so sensitive, companies don't want to talk about it. But there are a lot of techniques and approaches— cybersecurity is happening all the time," Hirsh says.
"The entire auto industry is working on security," says AT&T's Coursey. "By sharing threat information, we can increase safety and security for everybody. The industry is able to figure out how to detect threats and how to address them."
Additionally, according to a report from Strategy&, edge nodes should be able to detect suspicious behavior and to share this information so that potential security problems are averted early.
Privacy is an important concern for many users. How much data is shared? Who owns the data? Where is the data stored? How long is the data stored? Is the data encrypted?
GM's London stresses privacy and consent to receive services as a primary issue when working with a customer. An example is a new program enabling Amazon Prime members to have packages delivered to their cars. In order to activate this capability, the customer must "give consent to give the vehicle location, and to use locate, unlock and lock commands." Similarly, drivers' sharing data with insurance companies is a matter of mutual agreement. GM's London stresses privacy and consent to receive services as a primary issue when working with a customer. An example is a new program enabling Amazon Prime members to have packages delivered to their cars. In order to activate this capability, the customer must "give consent to give the vehicle location, and to use locate, unlock and lock commands." Similarly, drivers' sharing data with insurance companies is a matter of mutual agreement.
London continues: "All partner programs we offer require clear and express consent from our customer. Our customer makes a decision to opt in, and we are very clear about what data we will share with a partner." Hirsh believes that this sensitivity to customer privacy extends to all automakers: "I have the point of view that at the end of the day, vehicle OEMs will believe that protecting [customers' privacy] is most important."
Digital transformation has become essential for all industries. From grocers taking orders online to energy companies using sensors to make the electric grid smarter, to financial services firms customizing products and customer interactions, we've barely dipped a toe into the ocean of possibilities.
The auto industry is one of the sectors most actively engaged in exploiting new digital capabilities. Automotive manufacturers—now often referring to themselves as mobility companies—have adapted their businesses to capitalize on connectivity. From today's connected cars to tomorrow’s autonomous vehicles, these companies provide a glimpse into the future.
Related CIO Content