Analysis of the USDOT’s Regulatory Review for Self-Driving Cars (Part 2): Automated vehicle concepts
As discussed in the first part of this analysis, the USDOT Volpe Center’s review of federal regulations (i.e., the Federal Motor Vehicle Safety Standards, or FMVSS) for autonomous vehicles had two components: a “Driver Reference Scan,” which combed through the FMVSS to identify all references to human drivers; and an “Automated Vehicle Concepts Scan,” which examined which of the FMVSS would present regulatory obstacles for the manufacturers of autonomous vehicles. To perform this scan, the authors of the Volpe Center report identified thirteen separate types of “automated vehicle concepts” or designs, “ranging from near-term automated technologies (e.g., traffic jam assist) to fully automated vehicles that lack any mechanism for human operation.”
Here are those automated vehicle concepts as defined and described in the Volpe report:
Conventional Design Vehicles
1. Highway Automation – An optional system sold on new vehicles that offers human drivers extended periods of hands-free/foot-free driving on limited access highways. The system combines and integrates the functionality of existing lane-keeping, adaptive cruise control, and automatic emergency braking systems.
2. Driverless Valet – An optional system sold on new vehicles that allows a vehicle to park itself and retrieve passengers without a human driver seated behind the steering wheel of the vehicle. Upon arriving at a destination, the driver and any passengers could exit the vehicle and send the vehicle to locate a parking space. The vehicle finds and enters a vacant parking space, stops moving, shifts the transmission into park, and disables the propulsion system. When the driver is ready to retrieve the vehicle, he sends a signal to the vehicle using a smartphone app, at which point the vehicle enables its propulsion system, shifts into gear, and navigates to the driver’s location. During regular driving, the human driver remains in full control of the vehicle.
3. Truck Platooning – Optional equipment on a heavy duty truck allows the vehicle to engage in a close-formation platoon with other equipped vehicles. As a result, vehicles in the platoon reduce net drag and improve fuel economy. A human driver controls the lead vehicle and a combination of automation and human inputs control the following vehicles (e.g., a computer or artificial intelligence (AI) driver controls speed and following distance, and a human or AI driver controls steering). The lead vehicle broadcasts throttle, brake input, and vehicle speed data to trailing vehicles. Broadcast messages augment information supplied by onboard sensors.
4. Aftermarket Highly Automated Vehicle Kit – A third-party add-on system for existing production vehicles that provides host vehicles with highly automated capabilities. The concept system includes a roof-mounted module with necessary sensors and computer equipment to provide steering, throttle, and brake inputs through the vehicle’s controller area network (CAN) bus. As an aftermarket system, vehicles are equipped sometime after the original purchase of the vehicle.
5. Conventional Vehicle with Highly Automated OEM Add-on Kit – This concept is similar to the aftermarket kit, but represents a vehicle equipped at the time of sale with an OEM-provided add-on system to provide automation capabilities.
6. Highly Automated, Conventionally Designed Vehicle – This concept represents a vehicle that conforms with legacy design conventions generally assumed in the FMVSS, which is capable of fully-automated driving in most or all conditions. The vehicle can be operated from origin to destination with no direct input from a driver, or can be operated using conventional manual controls.
Advanced Design Vehicles
7. Highly Automated Vehicle with Advanced Design – The Advanced Design concept is capable of truly “driverless” operation and does not provide manual controls that would permit human driver operation. Not only does the vehicle not come equipped with a steering wheel, shifter, or pedals, but the design omits driver aids such as rear-view mirrors and cameras. The front seats can rotate 180 degrees. The vehicle retains a conventional windshield, equipment to aid in visibility (e.g., windshield wipers, defog/defrost, and exterior lighting), and telltales allowing the occupants of the vehicle (specifically, the person seated in the front left position) to monitor vehicle systems.
8. Highly Automated Vehicle with Novel Design – The ultimate incarnation of a driverless passenger vehicle, this concept omits any equipment that would otherwise be provided to allow for manual control or visibility outside the vehicle (mirrors, sun visors, windshield, windshield wipers, defog/defrost, headlights). Instead, the design emphasizes passenger comfort and convenience, providing a flexible, unconventional seating arrangement (making the designation of a “driver’s seat” ambiguous).
9. Riderless Delivery Motorcycle – A two or three-wheel vehicle with a small cargo compartment. The motorcycle operates without a rider along a programmed delivery route. The vehicle does not provide accommodation for driver control or seating and instead receives destination/route instructions remotely and operates using built-in automation capabilities.
10. Driverless Delivery Vehicle (Light Duty and Heavy Duty) – A four–wheel light- or heavy-duty vehicle designed exclusively for freight delivery. Like the riderless delivery motorcycle, this vehicle can navigate itself along a programmed delivery route without human control. The vehicle has no provisions for a human occupant or driver.
11. Low Speed Highly Automated Vehicle with Conventional Design – This concept is equivalent to the highly automated conventional vehicle described above, but is limited to 25 miles per hour.
12. Low Speed Highly Automated Vehicle with Advanced Design – This concept is equivalent to the highly automated vehicle with advanced design described above, but is limited to 25 miles per hour.
13. Low Speed Driverless Delivery Vehicle – This concept is equivalent to the light- and heavy-duty driverless delivery vehicle described above, but is limited to 25 miles per hour.
To be sure, we don’t hear much in the news about some of these types of vehicles, but all thirteen categories are important in terms of either economic impact or as a likely step on the road to widespread use of autonomous vehicles. The various technologies necessary for highway automation are already built in to many of the cars we drive and may well become standard within the next few years. Riderless delivery cars and motorcycles will appeal to courier-reliant industries such as law and medicine, as well as companies like Amazon seeking cost-effective methods for delivering consumer goods to customers. And automated low-speed vehicles would impact the jobs of an enormous number of Americans who work in industries such as construction and road maintenance.
The Volpe report opines that few of the FMVSS will pose a problem for the six categories of conventional-design vehicles. In other words, so long as a vehicle retains traditional motor vehicles’ basic design features–such as dashboard indicator lights, a steering wheel, headlights, etc–the addition of automated driving capabilities would not pose a regulatory compliance problem.
The report only identified two sets of standards that might prove problematic for such conventionally designed vehicles with self-driving capabilities–the foot-control brake rule and one of the theft and roll-away prevention rules. But I actually doubt that even these regulations would prove too problematic. The report suggests that driverless valet systems might run afoul of the theft/roll-away prevention rule that a car must be designed so that it cannot shift gears out of “park” unless the key is in the ignition. But as the report notes, an “electronic code” can substitute for a key and this has generally been interpreted as covering the sorts of keyless and push-to-start ignition systems that are already on the road. There is no reason to believe that courts or agencies would view an internet-controlled valet differently. And while the foot-control brake rule says that the service brakes “shall be activated by means of a foot control,” it does not suggest that alternative methods of brake activation cannot be added as long as the vehicle retains the capability for drivers to activate the service brakes by foot.
The FMVSS will be much more problematic, however, for autonomous vehicles that do away with the traditional features and mechanisms by which humans control motor vehicles. The report suggests that vehicles such as Google’s Car, which does not include a steering wheel or other mechanisms for human control, would run afoul of about a third of the FMVSS, including roughly half of the 100-series (crash avoidance) standards. As suggested in the entry covering the NHTSA’s widely misunderstood letter to Google, many of the standards would have to be rewritten–or an entirely new set of standards adopted–before such exclusively self-driving cars could hit the road.
Here too, the report often seems to read some of the regulations more broadly than is warranted. For example, it flags the rule that the parking brake “shall be operable by a person seated” in the front-left seat as potentially problematic for vehicles where the parking brake is activated and deactivated without human input. But as with the foot-control brake rule, nothing in the rule’s text suggests that the described method must be the only way to activate the brake.
That being said, there is little question that the FMVSS contains many rules that affirmatively require that a human be able to control the vehicle and/or be kept informed about the vehicle’s operational state. The absence of a steering wheel alone would violate numerous FMVSS–not just the requirements related to steering itself, but also regulations dependent n the presence of a steering wheel such as Standard 108’s rule that turn signals “must be self-canceling by steering wheel rotation.” And most of the FMVSS’ prescribed test procedures refer to the components of conventionally designed motor vehicles, and thus could not be conducted on vehicles that lack those features.
The bottom line is this: with vehicle automation technology advancing rapidly, there is a pressing need for the FMVSS to be amended or new standards to be adopted to account for the coming tide of autonomous vehicles. The Department of Transportation is well aware of this–as the Volpe Report shows–but curiously has been hesitant to initiate the rulemaking necessary to ensure that the law keeps pace with technology. Given the enticing convenience and safety benefits that self-driving cars would bring, that is a problem.