AV Policy 3.0: Analysis of Initial Comments about Automated Trucking
Commentators have expressed numerous concern about automated trucking.
Patrick Hickey raised the following concerns about automated trucking: (1) the need for a human in the cabin for inclement weather and difficult road conditions, (2) the greater need for maintenance in the automated trucking industry, (3) cybersecurity fears with automated trucks, (4) truck drivers were only responsible for 2.66% of truck deaths, and (5) the economic impact of lost trucking jobs.
Jerry Alford struck a harsher tone than Patrick Hickey about automated trucking and appeared skeptical about the ability of an automated truck to perform better than human truck drivers. Fred Perkins and Michelle Ardrey shared Jerry Alford’s skepticism about automated trucking.
A commentator named Concerned Driver also expressed concern about automated trucking. Concerned Driver initially incorrectly stated that numerous people have died from self-driving vehicles. However, only one person, Elaine Herzberg, has died in crashes involving highly automated vehicles. Concerned Driver’s theory for making our roads safer is to have stricter driving instruction and tests. Mark W. also suggested that better training will lead to safer roads than automated vehicles.
Weather and Difficult Road Conditions
Human drivers struggle to drive through hazardous weather. Approximately 21% of vehicle crashes, 19% of crash injuries, and 16% of crash fatalities are what the Federal Highway Traffic Safety Administration calls weather-related crashes–those crashes that occur in the presence of adverse weather and/or slick pavement conditions. Dissecting these numbers further, we see that approximately 70% of the weather-related crashes occur when the pavement is wet. 46% of the weather-related crashes occur while it is raining, and 18% of the weather-related crashes occur while it is snowing or sleeting.
Automated technologies currently struggle in inclement weather. But that likely will not always be the case. Manufacturers are already testing their technology in inclement weather and finding creative solutions to problems created by inclement weather. Automated technologies will get better at driving in inclement weather, though they will never be perfect in all weather conditions.
The commentators who complain about automated technology’s ability to drive through weather falsely assume that someone has to drive through inclement weather. If humans are bad at driving through inclement weather and automated technologies are bad at driving through inclement weather, then why would we want our trucks to operate through inclement weather? The solution to that problem is easy. No one should drive a truck during inclement weather. The solution is not to have a human, who also struggles in inclement weather, drive in place of the automated technologies.
This faulty assumption likely results because of the need for humans to drive through such weather conditions. As will be discussed further below, human truckers are regulated on their working hours and are paid by the mile. Thus, there is an economic incentive for human truckers to drive through inclement weather, so as not to lose precious driving time or miles. An automated truck can conceivably operate 24 hours a day, 7 days a week, stopping only to recharge its battery or to fuel up. Because the hour restrictions imposed by the Federal Motor Carrier Safety Administration (FMCSA) are intended to combat driver fatigue, these regulations will presumably not be necessary for automated trucks. Inclement weather will still affect the delivery schedules of automated trucks, but, at least in comparison to a human trucker, an automated truck can cease operating during periods of inclement weather and, given its unlimited ability to operate, be more efficient than a human trucker.
Maintenance in the Automated Trucking Industry
Automated trucks will need to be maintained mechanically and electronically. Although some of the comments indicate that these technicians will need to be in the truck itself, there does not appear to be an actual need for the Department of Transportation to require a human driver be in the vehicle. The technicians need not be employees of the trucking company. They may be third parties that service automated trucks within a given vicinity. The need for maintaining automated trucks may be a problem that industry will be able to resolve without the need for preemptive government oversight.
Cybersecurity Fears with Automated Trucks
Cybersecurity of automated trucks and all automated (and non-automated) vehicles is a serious concern for which manufacturers and governments need to prepare. Whitehat hackers have already shown that they can hack into connected vehicles and take control over aspects of the vehicles. Elon Musk has stated that one of the biggest concerns for automated vehicles is “someone achieving a fleet-wide hack.” Mary Barra, the CEO of General Motors, has said that protecting vehicles from cyber attacks is a “matter of public safety.” Manufacturers will need to ensure that their vehicles are safe from cyber attacks.
The National Highway Traffic Safety Administration (NHTSA) has already begun to study vehicle cybersecurity. According to NHTSA’s website, “NHTSA has adopted a multi-faceted research approach that leverages the National Institute of Standards and Technology Cybersecurity Framework and encourages industry to adopt practices that improve the cybersecurity posture of their vehicles in the United States. NHTSA’s goal is to collaborate with the automotive industry to proactively address vehicle cybersecurity challenges, and to continuously seek methods to mitigate associated safety risks.” NHTSA and the Department of Transportation need to continue researching cybersecurity and collaborating with industry on best practices to protect their vehicles.
Are Truckers Safe Drivers?
A recurring theme throughout many of the comments was how great truck drivers are at driving. Generally speaking, one would expect truckers to be better drivers than nontruckers because they have so much more driving experience than the average nontrucker. Americans drive approximately 17,600 minutes (~ 293 hours) and 10,900 miles each year. The average trucker is expected to drive 125,000 miles per year. Truckers drive more than the average driver and therefore should be more experienced. Additionally, we would expect truckers who drive poorly to lose their jobs and thus no longer be a trucker.
Conversely, truckers also drive for longer time periods and may drive more often than the average driver while fatigued. Most truckers are economically incentivized to drive as far as possible each day. They are paid by the mile. In order to combat trucker drowsiness and fatigue, the FMCSA regulates how long a trucker can drive before resting. The regulations limit a trucker to 14 on-duty hours (11 hours can be spent driving), followed by 10 hours of nonworking time. Because the economic incentives often outweighed the risk of violating those regulations, the FMCSA has, in recent years, required electronic monitoring of truck engines to ensure that truckers were not exceeding their driving restrictions.
Truckers also drive a more difficult vehicle to maneuver than the average driver. Driving a tractor-trailer is not easy. A tractor-trailer weighs up to 40 tons, has large blind spots (called “no zones” in the industry), and takes longer to brake than the average vehicle.
In summary, we would expect that the average trucker is a better driver than the average nontrucker, both based on experience and on the assumption that a trucking company would have otherwise weeded out a bad truck driver. Though the average trucker is probably a better driver than the average nontrucker, the average trucker likely drives more often with fatigue than the average nontrucker and has a greater economic incentivize to drive during through that fatigue and in inclement weather. A trucker also drives a more difficult vehicle than the average nontrucker.
With these assumptions about truck driving laid out, let’s look at the statistics for truck crashes. In 2016, there were approximately 475,000 police-reported crashes involving a truck. Of these crashes, 3,684 were fatal crashes and 104,000 were crashes that resulted in injury. The vast majority (83%) of the people who die in a truck crash are not the occupants of the truck itself. As a matter of physics, this result makes sense. Truckers usually have the largest vehicle on the road.
In 2016, the average fatality for million miles driven was 1.18 fatalities per million miles driven. The average fatality per million miles for a crash involving a truck was 1.46 fatalities per million miles driven. Neither of these statistics tells us what we really want to know, which is the percentage of crashes caused by truckers per million miles driven. I have not been able to find such a statistic to see if truckers truly cause more crashes than the average driver. These statistics, instead, tell us what we already know: everything else being equal, a crash involving a vehicle weighing up to 40 tons will result in more deaths than the average crash which, for most of the fatalities (89%), involve vehicles that weigh far less than a truck.
Will Automated Vehicles be Better than Human Truck Drivers?
Many people, including me, expect that automated vehicles will greatly reduce the number of car crashes. There are three general reasons why an automated driving system will be safer than human drivers. First, automated driving systems will not suffer from the main problems that underlie most crashes: human driver error. Human driver error crashes include those that result from distraction, intoxication, fatigue, drowsiness, and speeding. Second, automated driving systems will have better situational awareness and reaction times than the average driver. An automated driving system “‘sees’ everything in the vicinity; reacts at speeds humans cannot match; and constantly checks the performance of every component in the vehicle to ensure that it is functioning properly.” Third, over-the-air updates have the potential to mean that a given automated driving system will not cause the same crash more than once. Human drivers commit the same mistakes that other human drivers commit, and some human drivers routinely repeat their mistakes. An over-the-air update may mean that the automated driving system that caused the crash will not commit the same mistake and that the manufacturer’s other automated driving systems will not cause that same crash.
Automated trucks will have these same advantages over human truckers. When we look at the critical reasons for trucker crashes, the critical reason for 87% of crashes is the driver, 10% of crashes is the vehicle, and 3% of the crashes is the environment. The critical reasons for the drivers are 12% non-performance errors, 28% recognition errors, 38% decision errors, and 9% performance errors. Automated trucks should be able to greatly reduce these type of crashes because of their better situational awareness and decision making skills. Additionally, automated trucks may be able to reduce some of the crashes where the vehicle is the critical reason for the crash. Automated trucks will likely monitor all aspects of the vehicle and will not ignore maintenance issues and wear and tear. Automated trucks will also be able to monitor the behavior of other drivers and better prevent or avoid crashes that would be caused by human drivers. Accordingly, automated trucks should reduce the number of crashes caused by trucks.
Economic Impact of Lost Trucking Jobs
In 2016, there were approximately 1.5 million truckers with an average salary of $44,000. A person can be a trucker without a high school diploma or a college degree. As has been noted in an article entitled Self-Driving Trucks are Going to Hit Us Like a Human-Driven Truck, “truck driving is just about the last job in the country to provide a solid middle class salary without requiring a post-secondary degree. Truckers are essentially the last remnant of an increasingly impoverished population once gainfully employed in manufacturing before those middle income jobs were mostly all shipped overseas.”
Many fear that these employees will be displaced immediately by automated vehicles. However, this fear appears overstated in the short run for two principal reasons.
First, not all types of vehicle automation are capable of replacing the driver. The trucker will still be needed, from a technical standpoint, so long as the truck only utilizes lower levels (1-3) of vehicle automation. Additionally, level 4 automated driving systems are incapable of operating in all operational design domains. Thus, so long as a truck needs to operate outside that operational design domain, a driver will be needed to operate the truck.
Second, truckers do more than just drive their trucks; other tasks include “freight-handling, customer service, safety compliance, paperwork, and operating non-truck equipment.” A level 4 or level 5 automation system will replace only one facet of a trucker’s job, and until other aspects of a trucker’s job are replaced by automated features, there will be a need for a trucker in one capacity or another.
But even when automated trucks can fully replace truckers, mass displacement of truckers will not occur in a short time period. Instead, this process is foreseeable and will likely occur over a long period of time. The average age of a trucker today is 55 years old, and trucking companies are having difficulty recruiting young talent. At the end of 2017, there was a 51,000 truck driver job shortage in the United States. One of the reasons for the shortage in jobs is that young people fear that automated trucks will eventually displace them. Thus, the labor force is already preparing for potential displacement in the market for truckers.
All that is to say, truckers will not be replaced en mass at once. The process of displacement will be gradual, and the labor market has already and will continue to prepare for the displacement of these workers. Many of these workers may find that their job duties transitioned from driving tasks to secondary tasks, or that their jobs become more localized, such as driving trucks off highways in a specific region.
Governments, federal and state, should prepare for displacement of workers and prepare for the need to retrain workforces. Governments should not enact protectionist regulations to protect industries, including trucking from any potential displacement. Any such protectionist regulations come at a cost: either they delay automated trucking and cost society all of the expected benefits of automated trucking, or they limit those benefits by forcing industry to retain truckers in the trucks for purposes unrelated to safety.
The benefits of automated trucking, and the lives that can be saved, far outweigh any costs through loss employment. As noted above, truckers cause significant numbers of crashes each year–crashes that we expect automated trucks to be able to prevent. In addition to those benefits, automated trucking will provide all of the other benefits that we expect from automated vehicles, including reduced environmental costs and potentially reduced traffic congestion. Additionally, the average cost to ship dry goods is $1.85 per mile. Approximately, 26% of that cost is the driver’s salary and 39% is the cost of fuel. Many automated trucks are expected to be electric powered. Thus, there could be an over 50% reduction in costs. If automated trucks are significantly safer, the costs of insuring (4% of the total cost) these trucks may be reduced as well. With more streamlined braking and throttling, automated trucks should require less maintenance than human-driven trucks. Therefore, automated trucks have the potential to significantly reduce the cost of shipping goods in the country.
Second, forcing industry and consumers of that industry to subsidize displaced workers through jobs is inefficient. Society would be better off paying for retraining and providing unemployment and other benefits to these displaced workers than to require industry to keep truckers employed.
Automated trucking potentially creates a perfect storm for automated vehicle supporters. Those who will be economically harmed by automated trucks and otherwise fear automated vehicles more generally may scare the public about unmanned 40-ton vehicles crashing. Advocates of vehicle automation need to combat these fears but also have an understanding for persons who will likely be displaced by automated trucks.