The replacement to 2016’s “Future Automated Vehicles Policy” (FAVP) is called “Automated Driving Systems (ADS): A Vision for Safety 2.0.” And, as we started to discuss in my previous blog post, this policy builds on – but makes some serious reconsiderations of – the original policy. The “voluntary” aspect of the new policy was covered pretty well in the last post, so let’s not rehash old news. Instead, let’s look at what the new policy contains.

In a nutshell, the same reasons for having the new guidance exist as with the original guidance, because 9 out of 10 roadway crashes occur due to human behavior. Take the human out of the equation and the result, theoretically, is that you help reduce crashes, help reduce injuries, and help save lives. 

Along with safety, these technologies will help reduce congestion, enhance mobility, and improve overall productivity. It’s no wonder the feds would rather help speed development and deployment of these “wundertechs” so as not to impede progress (and, they don’t want the states to impede progress, either). 

That’s why, instead of four sections as in the original FAVP, there are only two. The sections eliminated include National Highway Traffic Safety Administration (NHTSA) regulatory tools, and the original agency request for Congress to help the agency gain new enforcement tools and resources. The two remaining sections cover “Voluntary Guidance” – the aspects all developers of ADS (Automated Driving Systems) need to keep in mind – and “The Federal and State Roles.”

A few points to start:

  • The new guidance is effective immediately, covering all classes of vehicles (Class 1-8) – including commercial vehicles. 
  • Unlike the original guidance, the new approach applies to the Society of Automotive Engineers (SAE) automated driving levels 3-5, not levels 2-5 as in the original policy. This does move the application closer to more automated systems, as level 3 automated systems conduct some part of the driving task and monitor the environment in some instances. However, the human driver needs to be present and alert, as the driver must take back control when the system requests.
    • You’ll recall that there is a lot of concern around level 3, especially re-engaging the human driver when needed. Will the driver be ready and able to engage as quickly as the system may require? Testing will tell, but a number of developers are skipping this and moving right to level 4. This action delivers a driverless experience – within constraints, however, or the operational design domain. You might be able to engage the driverless application on only a sunny day, for instance – the operational design domain for this particular application. But if you drive the system outside of this domain (sunny day conditions), you – not the system – will be driving the vehicle. So, as we’ve noted in past blogs, level 4 is really just a higher level of driver assistance.
  • Interestingly, NHTSA doesn’t overrule the Federal Motor Carrier Safety Administration (FMCSA) in the application of this technology for commercial vehicles. A trained commercial driver needs to be behind the wheel at all times, regardless of any automated driving technologies available on the Commercial Motor Vehicle (CMV), unless a petition for waiver or exemption has been granted. Truck testing and implementation of these systems will require special permission first. 

You may recall that the original FAVP had the Agency requesting that developers provide a safety assessment letter. The safety assessment was designed to be an executive-level readout of how the developer of the automated systems was going to cover 15 safety design elements in their approach.  The goal of the letter was to keep the Agency informed and help ensure that developers were thinking about some key elements in the development of their technologies.

Well, the new guidelines reduce the number of safety elements to 12, while doing away with the “request” and potential for requiring this via rulemaking – making the safety assessment letter completely at the discretion of developers. To my knowledge, as of the date of this blog post, only one developer, Waymo, has made any attempt to comply with the stated policy (which has been in place since September 2016).

Here’s how the elements compare between the current and original policies:

 

2016 Federal Automated Vehicles Policy

2017 ADS: A Vision for Safety 2.0

Operational design domain – How and where HAV is supposed to function and operate

Yes

Object and event detection and response (OEDR) – Perception and response functionality of the highly automated vehicle (HAV) system

OEDR

Fallback (minimal risk condition [MRC]) – Response and robustness of the HAV system upon system failure

Fallback (MRC)

Validation methods – Testing, validation, and verification of an HAV system

Validation methods

Registration and certification to NHTSA of an HAV system

 

Data recording and sharing: HAV system data recording for information sharing, knowledge building, and crash reconstruction

Data recording

Post-crash behavior – Process for how an HAV should perform after a crash and how automation functions can be restored

Post-crash behavior

Privacy – Privacy conditions and protection for users

 

System safety – Engineering safety practices to support reasonable system safety

System safety

Vehicle cybersecurity – Approaches to guard against vehicle hacking risks

Vehicle cybersecurity

Human machine interface (HMI) – Approaches for communicating information to the driver, occupant, and other road users

HMI

Crash worthiness – Protection of occupants in crash situations

Crash worthiness

Consumer education and training – (avoid Tesla Autopilot misinterpretations)

Consumer education and training

Ethical considerations – How vehicles are programmed to address conflict dilemmas on the road

 

Federal, state, and local laws – How vehicles are programmed to comply with all applicable traffic laws

Federal, state, and local laws

NHTSA does ask that entities share voluntary self-assessments with the Agency. The positioning of this “request” will, in my opinion, be kind of like Oliver asking for more porridge at the orphanage … and likely will be met with a negative response. 

Regarding federal and state roles: NHTSA still covers vehicles and technology, and retains rights over regulation on design and performance of vehicles. States, conversely, still cover human drivers, vehicle inspections, and rules of the road (vehicle operations). Because a number of states looked seriously at taking the previous policy and making it law within their borders, NHTSA recommends that states do not codify this guidance, hoping to help ensure that states don’t impede innovation and technological progress.

Of course, if the Agency is going to make everything voluntary and not provide regulations, states – such as California – may feel a need to encroach on the federal role to ensure the safety of their citizens. And of course, other states, like Arizona, may not.

To sum up the new policy:        

  • Voluntary, voluntary, voluntary
  • Please share with NHTSA
  • Do your best to make it safe because we can recall if we want
  • States – don’t tread on NHTSA’s authority

Don’t like it? Don’t worry. There are already rumors of an update planned for next year.

 

 

 

Bendix Blog

Technical and industry insight from OUR experts.

How We Got to Where We Are…

Driver assistance systems, they didn’t just appear overnight, right? This post discusses how antilock braking systems – or ABS – started us on the path to where we are today.

Read More

Not the Only Luddite in the Room

Autonomous, or driverless, vehicles are a long ways away. And more than just one person thinks so.

Read More

Getting to Safe Autonomous Vehicle Applications – NHTSAs “Federal Automated Vehicles Policy”

This blog takes a look at the “guidelines” released by the National Highway Traffic Safety Administration...

Read More