Security and wireless signal issues grow as vehicles become more connected.
Once again, a paradigm shift is upon us. Mobile connectivity has radically changed the automobile’s place in the world of connected everything. And that paradigm will shift even further once the IoE is in full bloom.
As all of this unfolds and technology marches on, some see the connectivity of the automobile as being a better and more powerful alternative to the smartphone. It is touted as offering everything from simple voice activated entertainment functionality to full autonomous operation, as well as infotainment, big data analytics, and artificial intelligence. Going forward, connected cars, as well as trains, boats, and planes, will have huge implications across a wide swath of the telematics arena.
According to the research firm Infonetics, connected cars are a prime driver for the M2M space, and will also tie closely to many of the sensors that will be part of the IoE. Gartner predicts there will be at least 250 million connected vehicles on the road by 2020. BI Intelligence, predicts that Internet connectivity will be available on at least 75% of all vehicles by 2020.
There are other encouraging numbers and trends. “For nearly every single feature tested, consumers prefer to access the feature through a built-in interface in the vehicle, without connection to a smartphone,” notes Jennifer Kent, director of research at Parks Associates. The only exceptions are browsing the Web, or accessing social networking apps.
Another Infonetics statistic presents the revenue derived by service providers from the connected car segment to be $16.9 billion from 2013 to 2018, for a CAGR of 25%. That is almost 21 times the expected growth rate of mobile voice and data services over the same time period. This is of extreme interest to the over-the-top players who view the connected car as fertile ground for push services.
OEMs, network operators and third-party vendors are taking notice. They are beginning to realize that if 75% of consumers consider connectivity a prime condition for purchasing their next car, they have a platform that can produce significant revenue for them.
While that may be a good thing, there are plenty of potholes along the way. Adding wireless and IP connectivity to a vehicle opens up business, technical, security and reliability issues.
Connected cars will use a plethora of wireless technologies. Everything from early 2G to upcoming 5G cellular platforms, various flavor of Wi-Fi, Bluetooth, multiple renditions of LTE, NFC, GPS/GLONASS, RADAR, DSRC, and emerging wireless platforms.
Some of these technologies are external, others internal. And they cover a myriad of frequencies. DSCR, for example, operates at about 5.9 GHz, Wi-Fi at 2.4 and 5 GHz, common cellular at 800/900 MHz, and 1.8/1.9 GHz. Collison avoidance radar runs at 24 GHz, 76 to 81 GHz, and 57 to 66 GHz. Then there is Bluetooth at 2.4 GHz, which is the same frequency as Wi-Fi, and there are known interference issues between them.
Once LTE and its variants become mainstream, they will operate at anywhere between 450 MHz and 2.1 GHz. And last but not least, while not directly part of the connected car spectrum, licensed frequencies in the ISM bands, (public safety, utilities, transportation) and other spectrum can have an effect on what is going on in and around the car if they get close enough.
Finally there are new frequencies being examined all the time for applications, especially for the IoE, and 5G. 3.5 GHz is one example. Others are 24, 28 and 38 GHZ and several others up to 120 GHz. It is still unclear what exactly these bands will be used for, and where, but the bottom line is that they will likely be considered for inclusion in all existing platforms at some point or another.
Individually, most of the propagation characteristics of these frequencies are well understood. Interference mitigation is also well understood. However, the cumulative effects of multiple-frequency mitigation pose significant integration challenges. In this ecosystem, where human lives can be on the line, it must be guaranteed that all of these wireless technologies work, individually, and together, as expected, and designed. That is a challenge.
“There are a lot of standards missing,” says Christoph Wagner, director of Rohde & Schwarz’ automotive market segment. Other standards are loosely crafted, even though they are for safety critical environments.
There are various types of RF interference, and many approaches to mitigation. Two of the more ominous are multipath and co-channel interference. Other types are PIM, OOBE EMI, and RFI. As well there is interference caused by physical objects (signal attenuation/bouncing).
The major issue with interference is degradation or corruption of the signal. It may not be a big deal if the in-car Bluetooth connection gets stepped on, but for things like collision avoidance, and self-driving communications, interference has to be eliminated. There certainly has to be more reliability in than today’s cellular networks, for example.
And, on the topic of security, unreliable wireless connections are an open invitation to hacking.
Interference mitigation techniques
Some technologies, such as spread spectrum, have inherent immunity to interference. The nature of spread spectrum makes it the top technology for both interference immunity and security. Spread spectrum is a well-understood technology and there is a plethora of data available on the topic.
Another promising area is smart antenna technology. Smart antenna systems (SAS) fall into three categories; SIMO, MISO, and MIMO. MIMO has attracted the most attention recently because it can not only eliminate the adverse effects of multipath propagation, but in some cases can turn it into an advantage.
There is an entire segment of the antenna industry focused on the design and development of SASes that is looking at how these devices can be applied to various platforms; vehicles and the IoE being high on that list. “Antennas are where the rubber meets the road in any wireless system,” says Jim Nevelle, CEO of antenna giant Kathrein-Werke. That puts them at the point of the connected part of the connected car design.
SASes are a result of new technology being applied to old hardware. The antenna itself is nothing more than a dumb radiation element, whether it is a 1000W tower antenna or a micro-sized chip antenna. The “smart” is in the antenna system, not the metal itself.
Today’s SASes are sophisticated enough to replace multiple antennas with one. For example, devices exist that can integrate Bluetooth, Wi-Fi, GSM, GPS, 3G multi-bands and 3.9/4G LTE into a single SAS.
There are several approaches to making the antenna smart. Essentially, smart antennas consist of a variety of transmission elements and DSPs. This allows the antenna system to gain the intelligence needed to “adapt.” That is why MIMO technology is so popular. Such antennas use sophisticated algorithms in the signal processing section to analyze and adjust or adapt to the varying multiple signal conditions.
One of the more bleeding-edge applications of SASes comes in the form of dynamic spectrum analysis (DSA). While DSA covers a broad range of areas, for this discussion it focuses on the intelligent analysis of signals, dynamically and in real time. The desired result is to improve spectrum efficiency via time- and space-dependent spectrum sharing among coexisting radio services, the exact environment of the connected car.
There are a number of proprietary technologies used to implement DSA. One uses algorithms to sense interference, and respond by dynamically shifting the transmission to the best available frequency. A derivative of that is using the algorithm to analyze the channel interference, but instead of changing to another channel, the DSA can analyze the signal with respect to the interference and continue to use the channel with degraded efficiency. This is particularly effective with MIMO systems, and results in much better link reliability.
SASes can use a multitude of approaches to mitigate interference and keep the reliability of the link as close to five-nines as possible. With the connected car, five-nines isn’t an option, it is a requirement. The more advanced the signal processing, the more reliable the link becomes.
Connected car security
One thing about the connected car is that the more connected it becomes, the more vulnerable it is to hacking, and that has been shown in several different high-profile breaches. And security it tantamount. If a phone gets hacked, no one is going to get killed. But if a car moving at 60mph gets hacked and the control is compromised, the results can be disastrous.
“Security has suddenly become a very hot topic – because of incidents like the Chrysler Jeep hack, and the Tesla Linux system hack,” says Andrew Patterson, Director Automotive Business Development at Mentor Graphics. “The reason for that is because suddenly more attack surfaces have emerged in cars. In the old days when it was just a wiring loom and there was a technology called the CAN bus, it was relatively hard to hack into that. You actually have to plug in a diagnostic connector, and to do that you have to get into the car or under the bonnet, then link the connector to something remote. It certainly couldn’t be done wirelessly very easily. But with the advent of connected cars, DAB (digital audio broadcasting) radios, Bluetooth links, keyless locking — all of those present new attack surfaces for potential hackers.”
There are many facets to connected car security, and part of that involves the system to which a car connects. Even if the car is secure, what happens if the traffic control system is hijacked?
Over-the-air updates are another challenge, and most of those revolve around software innovations, says Patterson. “It sort of started in the infotainment head units where people wanted new content rather like they do in a smartphone — you want an OS update — and carmakers have now realized they need to update their head units certainly more frequently than the lifetime of the car so they are designing in features to allow both plug in updates and over-the-air updates probably on a three- to six-month type of cycle. So if there are new apps, updated user interfaces, and in the last six months an important use case has been security patches. As people find software flaws they want to be able to issue patches into their cars.”
To do those updates securely, carmakers are starting to borrow from traditional computer networks, particularly with asymmetric keys. “You have a software patch, you encrypt it with a key that the carmaker would own, and then the owner of the car has a matching but different key, and that can only work together with the public key provided by the carmaker. Patches can then be delivered over the air, and they can only be unlocked by that particular user with their private key,” he says. “You can make it even more secure by introducing digital certificates, just as you would have for a Windows or Apple system on your laptop or desktop. In the future we may even see digital certificates from patch providers and within the vehicles so that send and receive know they are talking to trusted sources.”
Vehicle to vehicle (V2V) and vehicle to infrastructure (V2I) connections add more challenges. wrinkle in all of this. With V2V communications, there are a large number of signals being passed back and forth. The common number is 1,000 to 1,500 received messages per second from cars within a near proximity (roughly 100m). The most pressing issue is message authentication. Messages in cars will have to be verified in a very timely fashion. Imagine if there is a braking communication going between this car and the one in front. If the message verification takes seconds, the next thing the occupants of the car following might see would be the back seat of the car in front.
That is only one example, which can stratify across any number of V2V or V2I scenarios; lane changes, stop lights, emergency vehicles, etc. Expect that to be widely deployed within the connected car arena.
There are some solutions available, at least in concept. Figure 1 is an example of all of the attack surfaces the connected car will have.
One approach, used on the Lincoln Continental, and supplied by Cisco, uses a software-based security gateway that manages user identity, filters content and guards against security threats.
Another solution is virtualization, which can sandbox safety-critical components such as motive, steering and braking from components that are not safety-related, such as entertainment applications.
V2V and V2I are relatively recent developments, and there is little cohesion among the players, both hardware and software. Some solutions are in deployment, but most are still on the drawing board, especially with respect to the interconnect to the IoE. Vehicular security is a work in progress and still has a long way to go.
To create and develop secure cars and other road vehicles will be a rather arduous process. But on the brighter side, there is a lot of work going on, in all camps; auto manufacturers, security vendors and security hardware players. High-speed cryptography algorithms promise to bring solid security to the vehicular platform. Once they get some traction, expect things to ramp up much faster in the vehicular security arena.
Also expect to see new approaches to wireless vehicular technology that will make all the various wireless flavors play nice. This will be a challenge. There are many nuances in both the closed RF space in the vehicular interior, as well as the scope of the frequencies in V2V and V2I.
Overall, the amount of expected revenue that connected cars will generate is monumental. That has a way of bringing opposing forces and self-interests to the bargaining table. We’ll see just how many carrots will need to dangle to get that happening.
DSA – Dynamic Spectrum Access
DSRC – Dedicated short-range communications
EMI – Electromagnetic Interference
GLONASS – Global Navigation Satellite System
ISM – Industrial, Scientific, Medical
LTE – Long Term Evolution
MIMO– Multiple Input Multiple Output
MISO – Multiple Input Single Output
NFC – Near-field Communications
OOBE – Out of Band Emission
OTT – Over-the-Top
PIM – Passive Intermodulation
SAS – Smart Antenna Systems
SAT – Smart Antenna Technology
SIMO – Single Input Multiple Output
V2I – Vehicle to infrastructure
V2V – Vehicle to Vehicle