A Critique of the NHTSA "Levels" for robocars


Last year, the NHTSA released a document defining "levels" from 0 to 4 for self-driving technology. People are eager for taxonomy that lets them talk about the technology, so it's no surprise that use of the levels has caught on.

The problem is that they are misleading and probably won't match the actual progress of technology. That would be tolerable if it weren't for the fact that NHTSA itself made recommendations to states about how the levels should be treated in law, and states and others are already using the vocabulary in discussing regulations. Most disturbingly, NHTSA recommendations suggested states hold off on "level 4" in writing regulations for robocars -- effectively banning them until the long process of un-banning them can be done. There is a great danger the levels will turn into an official roadmap.

Because of this, it's worth understanding how the levels are already incorrect in the light of current and soon-to-be-released technology, and how they're likely to be a bad roadmap for the future.

Read A Critique of the NHTSA and SAE "Levels" for robocars.


Mr. Templeton provides a cogent analysis of the shortcomings of NHTSA and SAE Levels. His complaint is similar to Sierhus et al, those of the Defense Science Board, and my own: the NHTSA and SAE Levels do not account for varying use cases and operational environments. Brad worries that legislators will write bad laws based on The Levels, such as outlawing fully self-driving SAE Level 4 cars everywhere when they are safe in some situations, like low speed shuttle operations for campuses addressed by Navia's INDUCT.

Brad offers the following to complement The Levels:


If we do wish to have a taxonomy -- for discussion, not regulation -- that taxonomy should start by classifying road situations by their difficulty. Be advised there is no strict ordering, some situations are quite distinct from others.

Specially controlled roads with limitations on activity by other road users. For example, roads closed to the public, or to certain types of vehicles. Also includes roads with special infrastructure to assist the robocars. Roads meant for super-slow traffic. In other words, roads whose rules and structure are made for the robocar's capabilities.

Limited Access
Roads, like highways which are divided, well marked and allow only motor vehicles.

Low speed roads with limited traffic, controlled intersections (stop-signs and signals,) No blind corners, fast cross-traffic, or frequent schoolchildren. Driveways (if mapped.)

Traffic Jam
Situations where speeds are low, and cars are packed densely around, so it is only necessary to travel with the flow of the cars, and notice if that flow breaks up.

Higher speed roads, non-hard intersections, modest traffic, night-time low traffic periods. Entry-exit lanes on freeways. Unmapped driveways and parking lots.

Higher speed roads, most intersections, heavy traffic.

Unprotected high speed turns, highly congested merges, blind corners. Urban streets over 60km/h. Roads that differ from what was previously mapped. Construction zones. Jaywalkers.

Very Hard
Recent snow and ice. High speeds with dense pedestrian and cyclist traffic. Accidents or road blockages ahead. Pedestrians darting into traffic. Police redirecting traffic.

Brad also offers his own version of The Levels:


1: Hands Free (Cruise Control or other Level 2)
2: Eyes Free
3: Automatic (Unmanned, network connection required)
4: Autonomous (Empty, 100% networking not required)

Dear Brad
It's your thought on autonomous car's Levels. It has been gone 2yrs.
You mentioned that Level 4 is Empty, 100% networking not required.
If so, could we get level 4 in 10years?
Most car builders, IT companies are aiming their autonomous car launch in the early 2020s.
One of a city developer, I really hope that unmanned autonomous car or robo vehicle could drive safely & freely in my entire city.
It drives so many chainge in city scape & structure.
How do you think about the level 4? Is it real?

Add new comment