Published on September 4th, 2016 | by Michael Barnard0
NASA Is Mostly Wrong & Irrelevant About Tesla’s Autopilot
September 4th, 2016 by Michael Barnard
Recently, NASA Research Psychologist Stephen Casner told Scientific American that “Cars in 2017 equal airplanes in 1983.” This was part of an extended interview where he and others associated with NASA made a bunch of negative comments about Tesla’s Autopilot technology.
One has to wonder where they’ve been for the past 20 years of automotive history. And one has to wonder what data they are actually looking at.
To be clear, I’m a big fan of NASA and its work on aviation safety is unparalleled globally. It created a massive reporting database of all aviation incidents and makes it available to researchers, organizations, and companies globally to improve air safety. It’s been integral to making passenger airline travel the safest form of travel in the developed world:
“Excluding acts of suicide and terrorism, commercial aviation was the safest mode of travel in the United States, with 0.07 fatalities per billion passenger miles: ‘A person who took a 500-mile flight every single day for a year, would have a fatality risk of 1 in 85,000.’ “
But that’s airline passenger safety. Private aviation is still a very risky business.
“The majority of aviation fatalities that occur each year (85%) involved private aircraft (known as “general aviation”). On average, 549 people die each year in activities such as recreational flying (41% of flight hours), business travel (24%), and instruction (17%).”
And that’s the first place where NASA’s remarks fall short. Passenger cars are like private planes, not passenger jets. The cockpits are much, much simpler. The controls are much simpler. And the degree of risk if something goes wrong is much lower. While each traffic death is tragic, a person dying in a car collision is difficult to compare to a plane with 300 souls on board crashing from the sky.
If NASA had developed an equivalent road traffic safety database a few decades ago, it would see some very interesting things, things which for the most part don’t have significant comparisons to private planes and their advances.
- Handling — Being able to brake, accelerate and swerve to avoid potential collisions is important. Tesla is far from alone here and all cars in developing countries are much better than virtually any car of 50 years ago.
- Visibility — Good views from inside the vehicle and to the rear are important. There are still a lot of cars with poorer visibility, but rear cameras are becoming increasingly used and will become mandatory in some jurisdictions. Tesla is very good on visibility and rear cameras.
- Crumple zones — Tesla is the gold standard for this category. The front, side, and rear crumple zones in the Tesla Model S and X are enormous and very well engineered. This passive feature is saving the lives of Tesla owners in collisions, and it’s also reducing the impact on those in the other vehicles which hit Teslas, likely reducing severity of injuries there as well. It’s a feature that is only possible to this degree with electric cars, and all new cars need to shift from internal combustion to electric drivetrains over the next 20 to 30 years regardless.
- Airbags — Tesla has very good airbags. Lots of car companies do. This active safety feature has been saving people’s lives for a couple of decades.
- Seatbelts — Including seatbelts has been mandatory in developed countries and many developing countries for decades and saves countless lives. Shoulder belts save more. Five-point harnesses would save even more, but the cost vs benefit given other safety systems doesn’t seem to make them worthwhile.
- Traction control — Once again, this feature is becoming standard on new cars. It’s easier to implement and the results are better on electric cars, but this is still standard.
- Antilock braking — Ditto to traction control.
- Automatic emergency braking –– Lots of cars have this and it’s saving a lot of lives right now. Tesla has a very good implementation of it, but as the collision and fatality in Florida in May 2016 shows, it’s not perfect. It’s not clear if other manufacturers’ systems would be better or worse. Tesla is upping its game on this one right now and should soon be better than it has proven to be. Every car should have this feature and it should be better than Tesla’s implementation. This feature by itself is expected to reduce fatal crashes by 20% to 25%, and it’s directly on the continuum of features that Tesla is enabling.
- Driver assist features — This is where Tesla is truly head and shoulders above the current cars on the market from competitors. Different studies have found that Tesla’s ability to follow traffic in both speed and in lane are much better. And the evidence with about 50 million miles under Autopilot so far is that it’s very good. And Tesla’s statistics show that its cars under Autopilot have 50% fewer airbag deployments than its cars driven manually. However, Tesla is not satisfied with that and no one else should be either. Tesla is aiming for a billion miles under Autopilot before declaring it off Beta.
- Full autonomy — Fully autonomous cars will be even safer than the current level of Tesla’s Autopilot. They will have to be. Multiple companies are working toward full autonomy and it’s highly likely that this will be a feature on commercially available cars in two years, but regulatory regimes might not catch up for a while. Every major highway safety regulator is looking carefully at this, but the wheels of bureaucracy grind exceeding slow.
Cars have made tremendous advances in safety over the past several decades, ones which NASA seems to ignore in its comments and which aren’t particularly duplicated in private aircraft. Tesla is at the forefront of the next wave of automotive safety and per its statistics, which Tesla shares with the National Highway Traffic Safety Administration (NHTSA) — the governmental organization which actually does keep tabs on these things — Tesla is already seeing that previously mentioned 50% reduction in collisions involving airbags.
What exactly does the NHTSA say about Autopilot and other self-driving capabilities? Well, Mark Rosekind, the head administrator of NHTSA, stated recently that the auto industry and the use of Tesla’s Autopilot “can’t wait until it’s perfect.” The NHTSA is right, and it is the equivalent of NASA for the roads. It is the organization whose opinion is relevant and empirically based.
Tesla is incrementally, intelligently, and carefully adding more and more safety features to its cars. If NASA was paying full attention, it would be applauding it. So, NASA’s remarks should be taken with a grain of salt. It is asserting more relevance than it has in this domain.