Autonomous Vehicle Regulations – What You Need to Know
A few weeks ago, the National Highway Transportation Safety Administration (NHTSA) announced its initial policy related to autonomous vehicles. These regulations are intended to inform automotive and equipment manufacturers pursuing autonomous technologies about the items the federal government will be keeping a close eye on as autonomous vehicles take to the streets. As with most NHTSA policies, only some items are requirements, while others are guidelines, but manufacturers will be developing products with an eye toward adhering to all policies. Here’s what you need to know:
Federal policies differ from state policies
Manufacturers developing autonomous vehicle technologies need to adhere to federal, state, and local laws and regulations. This means that just following the NHTSA policy might not be enough. For example, California has its own set of regulations for autonomous vehicle testing and deployment that are in some ways more strict than the NHTSA policy. To date, 16 states have passed laws or defined regulations related to autonomous vehicle operations. For more detailed information on per-state regulations, see NCSL’s overview of autonomous/self-driving vehicles legislation.
“Autonomous” is a tricky term
NHTSA guidelines predominantly apply to vehicles classified as “Highly Autonomous”. This is based on the Society of Automotive Engineers definitions of automation levels as follows:
A “Highly Autonomous Vehicle” (HAV) is one which falls into level 3, 4, or 5 above. In simple summary, an HAV is a vehicle where either a human driver is not needed at all, or a human driver can give full control to an autonomous system under some conditions and not need to remain alert to driving activities. For real-world comparison, Tesla’s Autopilot system falls into category 2 above and is not technically considered an HAV under this definition. Google’s and Uber’s cars currently being tested fall into categories 3 and 4, and thus are considered HAVs.
It’s All About Operating Conditions
NHTSA makes it very clear that autonomous vehicles must be designed and launched with very specific operational parameters in mind. This is because it is assumed that early HAVs will not be fully capable of operating autonomously under in all environments and conditions. How manufacturers define the operating conditions for their vehicles will play a large part in how NHTSA evaluates their adherence to policy.
- Guidance for HAVs differs based on the defined operating conditions for the vehicle. NHTSA expects manufacturers to define the confluence of 4 factors:
- An HAV should be able to operate safely within the operating parameters for which it is designed and be able to handle all foreseeable events.
- In situations where the HAV is outside of its defined operating parameters or in which conditions dynamically change to fall outside of the HAV’s operating parameters (ex. If it starts snowing), the vehicle should safely be able to either leave the flow of traffic (ex. Pull to the shoulder) or transition control to a human driver. It is expected that the vehicle will clearly indicate to occupants and other vehicles/pedestrians that the transition is occurring.
- HAVs should be able to detect, deal with, and take preventative measures against a large variety of events within their operational parameters. For a list of all common events, refer to the NHTSA guidelines.
Safety Comes First
Regardless of autonomous technologies present, all autonomous vehicles are required to adhere to the same safety standards as non-autonomous vehicles. This includes things like seatbelts, airbags, anti-lock brakes, bumpers, etc. Additionally, NHTSA has placed safety concerns front and center in its policies in a number of key ways:
- HAVs must have extremely clear warnings and indicators when handing off control from the autonomous system to a human driver. As with many state policies, NHTSA also specifies that HAVs must be able to enter a safe state (e.g. come to a stop on the shoulder) if the vehicle has a problem or is unsure what to do and the driver does not take over immediately.
- NHTSA explicitly calls out that manufacturers should not assume human drivers in HAVs will be immediately willing and able to take over control in problematic situations.
- HAVs need to also clearly signal intent and status to entities outside the vehicle. For example, pedestrians and other vehicles on the road should be given visual indicators in order to mitigate potential accidents.
- NHTSA expects all HAV manufacturers to use sensors and algorithms to avoid accidents when possible and to minimize damage to people and property when an accident is unavoidable.
Data and Security Are a Big Deal
Large portions of the NHTSA policy are devoted to data collection, data handling, and cybersecurity. This should not be a surprise since data is at the core of autonomous technology, and security is an obvious weak point for any vehicle able to drive itself that also connects to the internet. Some key points of the policy to be aware of include the following:
- All data collection and sharing in conjunction with HAVs must be in accordance with manufacturers’ posted privacy policies and security agreements.
- Manufacturers should ensure that their privacy policies adhere to NHTSA’s (and the White House’s) guidelines related to Transparency, Choice, Respect for Context, Minimization and De-identification, Data Security, Integrity and Access, and Accountability.
- HAVs are expected to collect and report detailed information on all crashes and to make this information easily accessible to NHTSA for evaluation and aggregation..
- It is requested (but not required) that manufacturers make crash and positive outcome (i.e., near-miss) data available to other companies and researchers as well for the purpose of improving safety for everyone.
Education Is Important
NHTSA also makes it clear that simply releasing autonomous technologies to the public isn’t going to be enough. Manufacturers are expected to implement robust training and education programs for both employees and customers to ensure that they are aware of both the capabilities and deficiencies of their HAVs before ever getting behind the wheel. Suggestions for implementing such programs include the following:
- HAV manufacturers should make strong efforts to educate employees, dealers, marketing teams, and customers about the capabilities and limitations of their products.
- It is recommended that any HAV sales include hands-on driving time explaining exactly how to use HAV capabilities and what the driver still needs to be aware of.
- All vehicle manuals, specification labels, and training materials need to clearly explain how to use autonomous features and the conditions under which they should and should not be used by customers.
Less-Autonomous Vehicles Aren’t Off the Hook
Although most of the NHTSA policy is geared toward highly autonomous vehicles, NHTSA also specifically called out expectations for manufacturers regarding vehicles that can’t yet entirely drive themselves. This is no doubt intended to remind manufacturers like Tesla that safety expectations are still paramount even for “driver assist” features currently being marketed to customers. Specific call-outs include the following:
- More than half of NHTSA’s policy sections apply to driver assist features:
- NHTSA also specifically instructs manufacturers to be aware that partial autonomous driving features represent certain added risks since drivers may abuse the autonomous features or treat them as offering full autonomy when they do not.
As autonomous vehicle development rushes ahead, the regulations that govern them will be continually updated. Expect to see revisions to the NHTSA guidelines over the next year as well as more in-depth regulations on a per-state basis, especially as autonomous ride sharing picks up steam. However, with the Federal government now officially taking a stance, a major regulatory hurdle has been cleared for this exciting new technology.