Josh Joy is a PhD candidate in Computer Science at UCLA, where he focuses on providing privacy and resiliency for the Internet of Vehicles. Josh has a computer science background and worked at Toyota Info Technology Center, USA in summer 2017 researching connected vehicles. I recently interviewed Josh to learn more about connected vehicle technology, the Internet of Vehicles, and the impact of connected vehicle technology on self-driving cars.
What is a connected vehicle?
Traditionally, cars provide rudimentary sensors and data, such as gas level and oil changes. These sensors alert the driver to eventually take action. The driver, however, is the ultimate decision maker.
A connected vehicle, in contrast, has sensors inside and outside the vehicle. These sensors communicate data to the vehicle’s software system, to other vehicles (V2V), and potentially to surrounding infrastructure like roadways and traffic lights (V2X). The vehicle, in turn, makes decisions based on that data as to how to drive and where to go. Ultimately, human control is removed and the vehicle cooperates with other vehicles to maintain safety and deliver passengers in a timely manner.
Is a connected vehicle different than a self-driving car?
A connected vehicle may not be self-driving and vice-versa. A traditional vehicle can be connected. The implication is that a vehicle that is both connected and self-driving should be safer than a traditional vehicle with neither technology.
The purpose of connected vehicle technology is to enhance a car’s safety mechanisms via communication with pedestrians, infrastructure, and other vehicles. The U.S. Department of Transportation (U.S. DOT) estimates that connected vehicles can eliminate close to 80 percent of crashes involving non-impaired drivers.
For example, vehicles might be alerted about pedestrians crossing the street, infrastructure communication regarding pavement markings, traffic signs, and temporary traffic control, and vehicle-to-vehicle communication regarding sudden braking or intersection collision avoidance.
How do connected vehicles communicate with each other?
The leading technology for vehicles talking to each other relies on Dedicated Short Range Communications (DSRC), a type of WiFi for vehicles. In December 2016, U.S. DOT filed a Notice of Proposed Rulemaking to enable DSRC communication technology on all new light-duty vehicles. There are already ongoing deployment efforts in Japan and Europe to equip vehicles with DSRC.
There are competing technologies, such as Cellular V2X from Qualcomm. The difference between DSRC and Cellular V2X is that Cellular V2X adds the ability to talk to the cellular network.
That said, DSRC has been rigorously specified, tested, deployed since the early 2000s and is available today. Cellular V2X is still a recently new development.
Other promising communication mediums include visible light communication and millimeter wave communication. Both of these are more appropriate for platoon style fleets where vehicles communicate directly with the vehicle in front and behind. Millimeter wave has the potential to support the transfer of the large vehicle sensor data and the gigabit-per-second data rates required.
What are the problems with current V2V technology?
There are two pieces, technical and political.
First, V2V communication requires that senders and recipients have the same radio technology. Adding a radio to a single vehicle will not add any extra capabilities unless another vehicle also has the same exact radio. Even if a manufacturer deploys radios to its entire fleet, that would not guarantee the activation of collision avoidance as to vehicles made by a different manufacturer.
Thus, deployments are limited for now. Manufacturers must show that the currently-deployed systems effect some expected number of accident reduction. Naturally, the deployment process will be gradual.
This is where the political aspect comes into play. Everyone must somehow agree to use the same radio technology and commit to installing it in all vehicles on the road. However, there is a chasm between the carmakers and the cellular stakeholders. Naturally, the cellular stakeholders prefer vehicles to back haul data over their networks in order to generate further revenue. Carmakers, on the other hand, prefer to avoid the additional latency and expenses by leveraging their own vehicular cloud.
Government regulation has proposed the deployment of DSRC in all new light-duty vehicles because it is one of the first vehicular communication technologies developed, standardized, and tested. However, it remains to be seen whether the industry will commit to and massively deploy DSRC radios, pursuant to the proposed regulation.
How will connected vehicle technology impact self-driving cars?
Self-driving cars will not fully evolve until they can communicate with each other. The exponential increase of data capture and sharing in the age of connected vehicles (“Internet of Vehicles”) will overwhelm the current modes of wireless communication, meaning that data computation must move closer to the vehicles, as opposed to solely in the Internet cloud.
The Internet of Vehicles consists of vehicles communicating with each other as well as performing distributed processing among the vehicles (i.e., block chain processing), creating a “vehicle cloud.”
What is the end goal for connected vehicle technology?
Today’s trend is to place computation physically near the data source, as opposed to back hauling large amounts of data to the Internet cloud for processing. The result will be real-time analysis enabling “smart” everything, including buildings, energy grids, vehicles, and entire cities.
What do you see is the single biggest issue carmakers face in building connected vehicles for mass market use?
Repeatedly, history has shown that while communication networks provide exponential benefits, they also usher in a multitude of security and privacy threats. The Internet cloud is still trying to catch up with providing adequate security and privacy protection. There is a lot we do not understand, and we remain unable to confidently prevent cyber-crime, cyber wars, data breaches, espionage, and state-sponsored attacks even on our most “secure” systems.
Connected vehicles will realistically only be fully deployed once these same security and privacy concerns are addressed. Manufacturers’ primary focus has been on making self-driving vehicles “work,” while not paying close attention to the security and privacy issues. Given this rushed state of development, it is inevitable that the same cybersecurity issues that plague the Internet of Things will only become amplified in the Internet of Vehicles. ⊗
Autonomous vehicles are no longer just an idea. The technology is real and is, in fact, already on the road. Of course, cars have had autonomous features for a long time.
In 1958, the Chrysler Imperial became the first car to be fitted with the modern cruise control system, invented by a blind engineer named Ralph Teetor. In 1978, the Mercedes-Benz S-Class offered the first reliable anti-lock braking system, which uses wheel sensors to maintain steerability and prevent skidding when you hit the brakes.
Jump to 2015, when Tesla released Autopilot, an integrated system that uses camera, radar, ultrasonics, and GPS technology to help the Model S steer, change lanes, and manage speed on its own … er, mostly on its own.
Now, autonomous vehicles are a complete package. Manufacturers and lawmakers are gearing up for the imminent release of completely driverless vehicles (see my notes on California’s recently proposed AV regulations here).
Like any new ground breaking technology, i.e., the telephone, computer, or Internet, autonomous vehicles will have far-reaching effects on many aspects of society. No more stop-and-go commutes? No more distracted drivers blowing through stop signs? Yes, please. These are exciting times.
They may not be so exciting, however, for industries based on conventional cars.
Below, I discuss five groups of people who could face job uncertainty with the rise of autonomous vehicles. This could mean the loss of an entire job sector or a dramatic shift in what it means to work in a certain field. I also present some ideas on what we should do, if anything, to save these jobs.
#1 — DUI LAWYERS*
*Throw in personal injury lawyers and insurance lawyers who specialize in car accidents. More on that below…
Have you ever been arrested for drunk driving? Know someone who has? If so, you’re not alone. In 2014, over 1.1 million drivers were arrested for driving under the influence of alcohol or drugs.
Meet your neighborhood DUI lawyer, the guy you thought only lived on billboards and bus bench advertisements. He is one among thousands in a niche legal industry built to navigate the web of probations, fines, special licenses, and misdemeanor (or felony) convictions that will forever haunt your record. In California, your first misdemeanor DUI conviction could cost you over $15,000, about $2,500 of which includes attorney’s fees. Some DUI lawyers charge upwards of $10,000 depending on the complexity of your case, the time it will take to defend, and the lawyer’s track record, multiplied by your level of desperation.
When autonomous vehicles hit the road, will DUI lawyers still have jobs? DUI arrests will surely drop to zero, right?
The answer depends on the level of your car’s autonomy and your corresponding duty to monitor it while it shuttles you from the bar to the club.
A Level 5 vehicle, for instance, is fully self-driving for all situations and could come without a steering wheel and pedals (see Google’s Waymo). In this type of vehicle, those three dirty martinis should have no effect on your ability to slouch in the back seat, close your eyes, and let the car drive. You are, in effect, a passenger in a self-driving taxi (perhaps one of many in a self-driving fleet owned and operated by a ridesharing service).
Anything less than Level 5 may present drunk-driving issues. A Level 1 vehicle is what many of us now drive: a conventional car with some mode-specific features like cruise control that require constant human monitoring.
A Level 2 vehicle has more advanced driver assistance systems like blind spot monitoring, parking assist, and adaptive cruise control, but you still do all the driving. Both Level 1 and 2 vehicles are already on the road, and DUI lawyers have plenty of work to go around.
A Level 3 vehicle can steer, maintain proper speed, change lanes, and park itself, but only with the expectation that you will intervene at any moment. This mid-level of autonomy gives a false sense of safety, as we learned from the June 2016 fatal accident involving Tesla’s Autopilot. Coincidentally, some manufacturers like Ford and Waymo have altogether bypassed the development of Level 3 vehicles because their engineers have struggled to stay awake during long test runs, a big problem for cars that demand a driver’s situational awareness at all times. In other words, drunk driving a Level 3 vehicle will likely warrant a DUI arrest.
A Level 4 vehicle is highly autonomous within certain parameters, but may not be equipped to navigate extreme road conditions or venture outside specially mapped driving zones. Although a Level 4 vehicle can drive completely without human involvement, it still could have a steering wheel and pedals. This sounds like the ideal machine for those of us who want to auto-cruise through weekday commutes but retain the freedom to whip around tight corners on Sunday morning drives up Pacific Coast Highway. It also means that a drunk driver can choose to drive.
For those who have imbibed a margarita or two, the choice to drive or be driven should be an easy one. But that’s not necessarily the case. Ridesharing, for example, has not proven to be a panacea for the emboldened drunk.
A 2016 study conducted by scientists at USC and Oxford University concluded that the entry of Uber services into metropolitan areas has had “no aggregate association with the number of traffic fatalities.” According to the study, the average drunk person may not be sufficiently rational to shell out the extra cash for an Uber due to the low likelihood of a DUI arrest.
There are currently no scientific studies that demonstrate the reaction time of the handover of the self-driving vehicle to the human driver is sufficient or tolerable for the human driver to make a safe recovery. The influence of alcohol on the driver in a self-driving situation may accelerate or amplify the negative consequences of drinking while operating a self-driving scenario.
In sum, Level 4 vehicles might do little to curb drunk driving if they provide the option for humans to drive. This, of course, assumes that the same factors that keep drunk people from ordering an Uber (i.e., cost and low probability of DUI arrest) will also keep a drunk person from using autonomous mode in a Level 4 vehicle. It’s impossible to say whether this will be true.
But what if the autonomous vehicle isn’t driving fast enough for the drunken passenger’s liking? What if it fails to take his preferred route? What if he just wants to show off his driving skills to his drunk friends? Will the vehicle detect the passenger’s BAC and prevent him from taking the wheel?
The only way to truly eradicate drunk driving is a 100-percent Level-5 population of autonomous vehicles without steering wheels, pedals, or any means for humans to take control.
Remember those freaky automated capsules with rotating passenger compartments in Minority Report?
The National Highway Traffic Safety Administration estimates that 94 percent of car crashes in the U.S. can be attributed to human error (that includes texting and driving, among other things). In 2015, there were 6,296,000 police-reported car accidents in the U.S., 2,443,000 people injured, and 35,092 people killed.
A fully integrated grid of Level 5 autonomous vehicles — without steering wheels or pedals — will reduce human-caused accidents to nearly zero, even with drunk passengers. As a matter of public safety, we should root for this to happen.
DUI lawyers, in this case, will certainly be out of a job. Personal injury lawyers and insurance lawyers who specialize in car accidents will also have a hard time finding work.
So, what should we do to save the lawyers?
Luckily for them, they will still have licenses to practice law, which can go a long way. In other words, they’ll probably be just fine.
Notably, DUI lawyers also have a wealth of knowledge in Fourth Amendment search and seizure law (a common battleground in DUI cases). Armed with these tools, they may find work in cases where Level 5 autonomous vehicles are used to remotely shuttle illegal drugs through unsuspecting city streets. Moving on…
#2 — TRUCKERS
There are over 3.5 million truckers in the U.S. Those drivers move 10.5 billion tons of freight annually, which accounts for over 70 percent of the total freight on U.S. roads and upwards of $726 billion of annual gross freight revenues.
Many of those people could soon be out of a job. Among the potential uses for self-driving technology, long haul trucking is one of the key focus points for developers.
Indeed, a Class 8 truck (aka, an 18-wheeler) could soon drive from New York to San Francisco without taking a single rest break (an impossibility for human drivers due to federal rules that mandate 11-hour daily driving limits).
This will create efficiencies for the trucking industry and make our roads safer. In 2014, there was at least one large truck involved in 3,424 fatal crashes on U.S. roads, in addition to 82,000 injury crashes. Self-driving technology greatly reduces the probability of these kinds of accidents. Just watch this harrowing video of a Tesla predicting a car crash two vehicles ahead and engaging its emergency brake system:
Uber-owned startup Otto is leading the pack for autonomous trucks. On October 25, 2016, an Otto truck carried over 50,000 cans of Budweiser 120 miles from Fort Collins to Colorado Springs with no driver at the wheel. Other tech startups developing autonomous trucks include Embark, drive.ai, and Starsky Robotics. Mercedes has revealed plans for a long-haul truck called the Future Truck 2025, and Volvo is building its own model. Even Amazon is rumored to be investing in technology for an autonomous delivery truck after it issued a patent this year for a wireless control system that allocates reversible lanes depending on traffic flow.
The industry, it seems, is pushing hard to get humans out of the driver’s seat.
So, what should we do to save truckers? One answer is, maybe, nothing.
The trucking industry already has a severe driver shortage. A 2014 report by the American Transportation Research Institute concluded that the trucking industry “has an aging employee base with a shrinking replacement population of younger trucking industry workers.” The average age of truckers is 47, a jump up from the median age of the total U.S. labor force, 42.2. Meanwhile, truckers earn a little over $40,000, about $8,000 less than the national average of all occupations.
In other words, truckers tend to be older and earn less than the average American worker. In addition, truckers often spend weeks away from home, long hours alone on the road, and face increased rates of obesity, high blood pressure, and type 2 diabetes.
In the long run, then, fully autonomous trucks will replace a working population that already suffers from low pay and detrimental health issues inherent to the job. From a public policy perspective, this could be viewed as a good thing.
Autonomous trucks will also save the trucking industry money. According to a 2013 Morgan Stanley report, autonomous trucks will save the industry over $70 billion annually in labor costs, $35 billion in fuel efficiency gains, $27 billion in productivity gains, and $36 billion in accident savings.
In the short term, truckers will not necessarily lose their jobs en masse. Lawmakers will require licensed drivers to remain behind the wheel, at least in the early stages of self-driving technology. Humans may still need to steer through tough road conditions, execute safety maneuvers in emergency situations, and navigate three-point turns in congested city streets.
The job, however, will be different. Rather than long hours focusing on the road, drivers will have more supervisory roles monitoring computer systems and taking control when necessary. This kind of role may require a different skill set and probably a special license. When the first autonomous trucks arrive, lawmakers should consider subsidizing the additional training needed to obtain self-driving licenses (paid for by taxing companies for the use of self-driving technology). With careful planning, the gradual transition to fully autonomous trucks need not signal the sudden demise of an entire segment of the American workforce.
Long term, however, human drivers will be phased out even for 18-wheelers. Self-driving trucks will simply be less costly, more efficient, and safer than humans.
When that time comes, lawmakers should consider extending additional retirement benefits to lifelong truckers in the form of tax deductions or credits.
#3 —METER MAIDS & VALETS (BASICALLY, ANYTHING RELATED TO PARKING)
Perhaps the most exciting prospect of autonomous vehicles is eliminating the need to park. Many developers are devoting their efforts to fleet-based services, which means you’ll be able to request an autonomous vehicle on your phone and hop out at your destination just like Uber or Lyft. An autonomous vehicle could run 24/7 and rarely leave the road, even for a charge. With the growth of magnetic induction technology, cables built into roadways could charge autonomous vehicles as they drive.
For parking-based labor forces, this is not great news.
First, meter maids will have fewer tickets to write as autonomous vehicles multiply. Consumers will opt to use ridesharing services like Uber rather than pay the costs of owning, maintaining, and insuring a car, not to mention suffer the inconveniences of finding parking, paying for it, and adhering to arbitrary time limits.
Ultimately, meter maid jobs could become obsolete.
But there’s a twist.
Cities really want to keep their meter maids. San Francisco makes over $100 million per year in parking citation revenue. Los Angeles makes $165 million per year. In 2015, New York City made $565 million from parking tickets.
Cities use revenues from parking as the foundation of their annual budgets, which include the salaries of city employees who issue the citations, i.e., meter maids.
If meter maids are gone, how will cities make up for these lost revenues?
One answer: taxes! By taxing self-driving services for the use of city streets, cities can generate the income necessary to install and maintain infrastructure to support those services, such as magnetic induction charging lanes, computer-choreographed intersections, cyber-hacking defense systems, and additional cell towers to sustain the massive data streams that will flow from autonomous vehicles to their developers. But will this be enough to make up for all that lost parking revenue?
Who knows? Besides, nobody likes taxes.
There’s an alternative: self-driving technology presents other opportunities for citation revenue. For example, if a passenger were to navigate his autonomous vehicle outside a designated safe zone, will there be a $63 fine? What if a passenger drives in the charging lane when he doesn’t need a charge? Will that be akin to a $491 carpool violation?
The answer is probably yes.
But an important question remains — will these citations be delivered by a meter maid? Or will the autonomous vehicle self-report your violation? (This is more desirable for local governments looking to make a buck … and less desirable for those of us who want to retain some sense of privacy).
Either way, assuming a world filled with fleets of Level 5 vehicles, meter maids will be out of a job or doing something much different.
Parking attendants and valets will also eventually become less necessary, though this could take many years. To the extent that the transition to fully autonomous vehicles is a gradual one, valets will be in demand as long as consumers choose to drive their Level 1 and 2 cars.
#4 — CAR DEALERS
Car dealerships first arose in the early 1900’s so that auto manufacturers could focus on building good cars while local salesmen could educate and sell the cars to local communities. In the 1950’s, state legislatures across the country enacted dealer franchise laws to protect car dealerships from abusive practices of manufacturers. These laws prohibited manufacturers from owning dealerships or selling cars directly to consumers. At the time, legislatures sought to protect dealers — typically local, family-owned businesses — who spent money and time growing a local market only to have a manufacturer open a rival dealership and undercut the dealer’s prices.
The dealer franchise model has been a thing of contention among manufacturers, dealers, and consumers for many years, and Tesla recently uncovered a chink in its armor.
In the late 2000’s, when Tesla began to sell its high performance cars directly to consumers, dealerships were furious. Elon Musk argued that Tesla did not have any existing franchises in place, so there were no unfair competition issues raised by selling directly to consumers. Lawsuits commenced, and now some states allow Tesla’s direct sales while others do not.
Autonomous vehicles have the potential to eliminate this patchwork of confusing state laws. Specifically, fleets of autonomous vehicles can dispose of the franchise sales model for three reasons.
First, manufacturers will sell cars to fleet services or entire cities. For example, if Uber wants to deploy 100 self-driving cars to serve San Francisco, Volvo will build and sell those cars just to Uber. This way, Volvo will keep its costs down by building exactly as many autonomous vehicles as it knows it will sell, and Uber can establish a uniform system of autonomous vehicles that all speak the same language and execute the same functions (uniformity of communication is a highly desirable trait for a safe self-driving future).
Second, individual consumers won’t own cars. Why own or lease a car only to worry about maintenance, insurance, and other costs? When self-driving fleets become ubiquitous, consumers will choose ridesharing services over car ownership.
Third, repairs and maintenance of self-driving cars will be infrequent and highly specialized. Dealerships currently make narrow margins on new car sales and sometimes even lose money. In contrast, they make 44 percent of their gross profits from service and parts replacement.
Self-driving cars will mostly be electric (hybrid in some cases). Electric vehicles outperform internal combustion engines in nearly every aspect of car maintenance. There are fewer moving parts, no need to change the oil or transmission fluid, clean the filter, replace a spark plug, or repair the exhaust system. In addition, self-driving technology reduces the probability of erratic driving behavior and body damage due to accidents. Software updates can be pushed to cars via Wi-Fi. Any computer glitches than cannot be fixed remotely likely need the expertise of an engineer at the manufacturer’s own facility.
Dealership service departments will, in turn, have fewer repairs to make. This means less revenue to subsidize competitive sales prices for new cars. Thus, by making better, smarter cars, manufacturers will knock out dealers’ cash cow (service departments), and make the dealership sales model unsustainable.
Naturally, there are some big assumptions here, one being that fleets of Level 5 autonomous vehicles will be universally available to the point where it is less costly to pay for a ridesharing service than to own or lease a car. This certainly won’t be the case right away, and it might not happen for many years. But when it does…
What should dealers do? One thing they have going for them is that, collectively, they own a lot of real estate. That real estate can be sold or converted to other uses. Some dealers might survive by virtue of a niche market for Level 1 and 2 cars.
Another possibility is to revamp service centers to accommodate Level 4 and 5 cars, hire expert EV technicians, and cut deals with manufacturers to acquire exclusive rights to service their self-driving cars.
Whatever the solution, it could eventually end the 100+ year franchise model as we know it.
#5 —TAXI DRIVERS (AND UBER & LYFT DRIVERS)
This is the most obvious endangered job on the list. According to the Department of Labor, there were 233,700 taxi drivers in the U.S. in 2014. That includes 14,700 taxi drivers in New York, 13,910 in California, and 12,040 in Nevada, the top three states for taxi employment. Uber had more than 160,000 drivers nationwide, and Lyft had over 100,000.
The taxi lobby has staunchly resisted Uber and Lyft, especially in big cities where taxi drivers often face stiff regulations and licensing fees that ridesharing services can largely ignore. Soon, these arguments will be moot.
Autonomous vehicles will stamp out the need for human drivers altogether.
Exactly when this transition will occur is hard to say, but many manufacturers aim to release fully autonomous vehicles within the next five years.
Tesla has announced that all its new cars are currently being manufactured with Level 5 autonomous hardware. Ford hopes to release a fleet of self-driving cars in 2021.
The release dates of self-driving cars are not necessarily the same dates that taxi drivers will become obsolete. Like any other new technology, humans must learn to use autonomous vehicles and accept them as safe, reliable, and cost-effective modes of transportation. Until that happens, human drivers will persist.
Cities and states generate big revenue from parking citations and traffic tickets. San Francisco makes over $100 million a year from parking citations alone. Los Angeles makes $165 million. New York City makes $565 million.
When Level 5 autonomous vehicles (AV’s) hit the road, this revenue could shrink dramatically. In a world where fleets of law-abiding AV’s shuttle passengers around, there will be far less human error, and thus, fewer citation opportunities. That means less money to fund state and local programs like police officer training, wildlife protection, emergency medical services, public improvement projects, and court operations.
As lawmakers write new AV laws, they should consider how to replace this lost future revenue. One option is a road use tax.
Massachusetts’ Proposed Road Use Tax
In Massachusetts, Rep. Tricia Farley-Bouvier and Sen. Jason Lewis recently introduced Bill S.1945, which imposes a “road usage charge” for AV’s (you can read the full text of the bill here). Here are a few of the bill’s key provisions:
2.5 cents per-mile base rate. The bill proposes a “base per-mile rate on autonomous vehicles of no less than 2.5 cents per mile.” The rationale for “2.5 cents” is not stated, but this number could change before the bill is finalized. On its face, it appears to be a palatable number for the average consumer that could still generate substantial revenue for the state. The average U.S. driver travels over 13,000 miles a year. Based on a 2.5-cent rate, that would amount to a $325 road usage bill.
Reduced rates for passengers. The bill proposes a rate reduction “for each passenger … per mile.” In other words, the bill encourages carpooling. The more passengers per vehicle, the lower the rate per mile for that vehicle.
Increased rates for “zombie cars.” There is an increased rate “for each mile traveled without a passenger.” So, the bill discourages the traffic inefficiencies caused by AV’s occupying road space without carrying any human passengers, a phenomenon Senator Lewis calls, “zombie cars.”
This seems reasonable, but there is currently no exemption for an empty ridesharing AV on its way to pick up one or more passengers. Different usage rates based only on passengers physically inside an AV may not truly reflect whether the vehicle is performing a valuable service or idly taking up road space.
Reduced rates for off-peak hours. Higher rates will apply from 8:00 am to 8:00 pm in “severe congestion zone[s].” On the other hand, there will be reduced rates during “off-peak travel hours.”
Reduced rates for low-income passengers and low-access areas. The bill proposes reduced rates for AV passengers “whose personal income, as documented by tax returns or other credible evidence, falls below a threshold established by regulation.” There are also reduced rates for AV’s driving “in specified geographic areas where no or few public transit options are available.”
These provisions aim to lessen the burden of the road usage charge on low-income individuals. They fail, however, to set forth guidelines on how lower rates will be assigned. Will there be an application process with a reviewing agency? Or will lower rates automatically apply based on an individual’s tax returns? In addition, what qualifies as “other credible evidence”? These questions should be considered before the bill is finalized.
Data collection. Each AV must capture and store data including “real-time distance traveled and real-time number of passengers.” This data must be stored for up to 18 months in a manner that can be “cross-referenced” by the Massachusetts Department of Transportation (MDOT).
Privacy concerns. AV passengers will naturally be hesitant to self-report information about themselves and their mileage statistics to local authorities. Under the proposed rules, MDOT must be able to “cross-reference” passenger data, i.e., access it independently without the passenger’s consent.
This level of government access to passenger data has been fiercely opposed by privacy advocates in relation to conventional cars. It remains to be seen whether it will be deemed tolerable for the purposes of collecting a road use tax for AV’s.
Alternative to a Road Use Tax Based on Mileage & Passenger Data: Vehicle Categories
Mileage and passenger data may generate precise usage rates, but this system presents difficult privacy issues. It also places a high burden on manufacturers to install systems that accurately capture that data and ensure that it is not hacked, leaked, or corrupted.
Instead, lawmakers should consider a road use tax for AV’s based on vehicle categories. For example —
Category 1: Consumer AV less than 4,000 lbs. / Tax: $100
Category 2: Consumer AV more than 4,000 lbs. / Tax: $200
Category 3: Ridesharing AV / Tax: $300
Category 4: Long-haul AV truck / Tax: $400
These numbers are admittedly rough. The point is that a ridesharing AV can be expected to drive more miles than an individually-owned consumer AV. Long-haul AV trucks can be expected to drive even greater distances. A road use tax could be imposed on AV owners based on these broad vehicle categories (and adjusted based on how lawmakers want to incentivize the use of each kind of vehicle).
A category-based approach might overtax and undertax certain users. On the other hand, it reduces the need for independent government access to passenger data. It also sets up a more predictable tax rate for AV owners and a more predictable income stream for cities and states.
Ultimately, whichever tax scheme prevails could significantly impact the way manufacturers design and deploy AV’s.
You can track new state-by-state AV legislation here.
The proposed regulations are an important step forward for manufacturers who now have a more specific set of guidelines to test and deploy autonomous vehicles. By clarifying the requirements for autonomous testing and deployment, the proposed regulations increase the certainty with which manufacturers can develop self-driving technology.
Still, some important questions remain unanswered and should be addressed by policymakers at the April 25, 2017 hearing in Sacramento to finalize the new rules.
Autonomous Mode vs. Conventional Mode: A Liability Question
Previous drafts of the regulations defined “autonomous mode” as a type of “vehicle” that is “driven without active physical control by a natural person sitting in the vehicle’s driver’s seat.”
The new regulations define “autonomous mode” as,
the status of vehicle operation where technology that is a combination of hardware and software, both remote and on-board, performs the dynamic driving task, with or without a natural person actively monitoring the driving environment. § 227.02(a).
“Conventional mode” is defined as,
the vehicle is under the active physical control of a natural person sitting in the driver’s seat operating or driving the vehicle with the autonomous technology disengaged. § 227.02(d).
By defining “autonomous mode” as a status, the regulations address an important aspect of self-driving technology: autonomous vehicles could come packaged with two modes — “autonomous” or “conventional.” As such, a human driver could disengage autonomous mode and take control with a steering wheel and set of pedals.
A dual-mode autonomous vehicle is appealing to consumers who might want the freedom to drive at their discretion. It is also appealing to manufacturers who can sell cars capable of driving beyond specially designated driving zones with human assistance.
It also presents interesting questions of liability:
In the event of an accident in autonomous mode, could the human driver have taken control? If so, how much liability should we place on the driver?
A fully autonomous vehicle, on the other hand, may be more limited in its driving range and capabilities, at least in the early stages of self-driving technology. For example, if a tree branch falls on the road, will the fully autonomous vehicle know to cross the double yellow line to get around it, or will it sit in the lane and wait for someone to clear the obstruction? Will it practice defensive driving, i.e., swerving into the shoulder when the car next to you changes lanes without looking?
Questions of liability, however, are “easier” for the fully autonomous vehicle in the sense that liability should be placed on the manufacturer when the human has virtually no control of any dynamic driving tasks. This is additional motivation for manufacturers to produce cars that come with a “conventional mode,” so that they can avoid this obvious presumption of liability when accidents occur.
Passengers Cannot Pay for a Test Ride
A “passenger” may “summon a vehicle or input a destination,” but does not monitor the vehicle’s dynamic driving tasks, like steering, acceleration, changing lanes, or parking. Notably, the definition of “passenger” also includes the following:
A member of the public may ride as a passenger in an autonomous test vehicle if there are no fees charged to the passenger or compensation received by the manufacturer. § 227.02 (j).
This appears to be a direct jab at Uber, which attempted a rather early release of its autonomous car service in San Francisco in December 2016. In addition to lacking the necessary permits, the Uber vehicles reportedly ran at least six red lights during the short-lived experiment. For now, it appears the DMV hopes to discourage autonomous ridesharing until the technology has been fully vetted.
For those of you who know someone working to develop autonomous vehicles, here is your chance for a free ride. Manufacturers cannot charge you for riding as a passenger in their test vehicles.
But don’t get too excited. Manufacturers must comply with a series of stringent requirements to test their vehicles, which includes driving them within predefined zones and notifying local authorities of their whereabouts. In other words, free-ranging rideshare services are not yet available for the general population.
Mandatory Requirements to Test Driverless Vehicles: Who is “At Fault”?
Among the many conditions required to obtain a testing permit for driverless vehicles, manufacturers must do the following:
Hold a minimum $5 million insurance policy;
Coordinate testing locations and conditions with local authorities;
Instruct law enforcement on how to interact with the test vehicle and safely remove it from the roadway in case of emergency, and
Provide a two-way communication link between any passengers and a remote operator who must continuously monitor the car’s status.
Further, the manufacturer must agree to assume complete liability for any damage caused “to the extent that the autonomous vehicle is at-fault in any collision.”
This seems to leave a fair amount of wiggle room for the manufacturer. For example, an autonomous vehicle that suddenly brakes to avoid an obstruction could cause a rear-end collision, whereas a human driver would otherwise brake less suddenly to reduce the risk of such a collision. In this case, is the driver following the autonomous vehicle completely at fault? Or should the manufacturer assume some responsibility for programming overly rigid braking rules?
What if a child jumps into the middle of the road, and the autonomous vehicle swerves and hits a dog in a crosswalk? Is the autonomous vehicle “at-fault” for the dog’s injuries? Should the manufacturer have written “safer” code to prevent this type of accident, and how can we possibly determine that?
There should be mechanisms in place to determine fault when accidents arise. But there’s a problem:
Determining fault may require access to an autonomous vehicle’s proprietary machine learning algorithms, which manufacturers have valid reasons to protect.
Lawmakers should consider when and how to mandate access to proprietary data and work with manufacturers to establish a clear set of rules. By doing so now, they will save a lot of time, energy, and money when thousands of autonomous vehicles rule the streets.
What About Cyber?
The regulations are missing at least one important element: testing for cyber-attacks. Cars with computers can be hacked, and it’s not hard to do. In 2015, two hackers remotely turned off a Jeep while it was cruising on a highway, and that was not an autonomous vehicle.
Autonomous vehicles will be more reliant on computer systems than ever. That makes them more vulnerable to cyber-attacks than conventional cars. Further, a cyber-attack on an autonomous vehicle could be a lot worse. Because autonomous vehicles must communicate with each other using vehicle-to-vehicle technology, an attack on one car could compromise an entire fleet of 100 autonomous vehicles.
A certification that the autonomous vehicles have self-diagnostic capabilities that meet current industry best practices for detecting and responding to cyber-attacks, unauthorized intrusions, and false or spurious messages or vehicle control commands. § 228.06 (a)(9).
Industry best practices for cyber-safety have been published by the NHTSA (National Highway Transportation Safety Administration) and Auto-ISAC (Automotive Information Sharing and Analysis Center).
Self-certification may not be enough, especially because manufacturers are rushing to be the first to hit the road with a fully autonomous vehicle. Lawmakers should work with manufacturers to devise a rigorous series of tests which must be satisfied before a Deployment Permit is issued. This may seem burdensome, but it is a necessary cost to ensure a safe and happy public introduction to this wonderful new technology.