Global Bioenergies obtains a €400K grant from BMBF to produce renewable gasoline additives; Audi to use for engine testing
ISO/TS 19880:2016 – New technical document for hydrogen stations

Driver video: Tesla Model S adaptive cruise control rear-ends stopped van

A Tesla owner in Europe has posted a YouTube video of his Model S rear-ending a stopped van.

In the notes accompanying the post, while asserting the superiority of the Model S, he outlined the safety-systems failures:

  1. The TACC, active cruise control did not brake as it normally does.

  2. The automatic braking system (AEB) did not make an emergency brake.

  3. The forward collision warning turned on way too late, it was set to normal warning distance.

  4. The TACC actually was speeding up just before he did hit the brakes.

Yes, I could have reacted sooner, but when the car slows down correctly 1,000 times, you trust it to do it the next time to. My bad…

I was looking right to find a spot to merge, and did not realize soon enough that the TACC was accelerating towards the van. So my reaction was too slow because of that. It takes a second to realize that the system is failing and you have to take over. In normal operation, the AP slows down as soon as another car puts one wheel on the lines to your lane.

Comments

electric-car-insider.com

This video has been up for a while, nonetheless, shows the importance of maintaining situational awareness and readiness to resume manual control at all times.

While we're chatting at Superchargers, I've begun asking fellow Tesla drivers about autopilot use. A majority report inattention and being lulled into a sense of complacency. Most say they have changed their mode of AP driving after hearing of Josh Brown's death.

Tesla needs to step up the communication and training of AP drivers. The last guy I talked to received only 10 mins orientation when picking up his car (at his request - had more important things to do apparently, than learning about a sophisticated control system his life - and his children's lives - was going to depend on).

@Tesla Motors - there's an education systems failure that needs to be addressed.

Brian Petersen

Further evidence that Tesla's system is not ready for prime time and hasn't gone through sufficient validation testing.

Lulling people into a false sense of security is extremely dangerous.

Account Deleted

E.C.I. any concrete suggestions for what more Tesla should do? I personally think Tesla has made it crystal clear that the responsible driver is you and that the AP is a conveniences feature that is incapable of autonomous driving so it needs nearly constant oversight. Maybe Tesla should offer 30 minutes weekly courses at their showrooms about how to best make use of the AP and maintain their Tesla. However, I don’t think it will make much of a difference.

The important thing IMO is that Tesla can document fewer accidents with autopilot on than without autopilot on. Tesla has said that they already has enough statistically evidence to conclude that using autopilot is much safer than not using autopilot. They measure that by the number of airbag deployment which is much less per mile driven when AP is activated. In 2 or 3 years Tesla will have solved the problem entirely by developing a fully autonomous system so that the car can drive itself at all times everywhere. All it takes is a better sensor package and some more programming. Model 3 will get the needed sensor package and processing powers when it arrives but I expect Model S and X will get it first. Subsequent software updates will make the cars fully autonomous by 2019. I also expect Tesla will launch a driverless taxi service in about 2020 with a new tandem two seater BEV with falcon doors in 2020 and no steering wheel or gas pedal. It will be rolled out everywhere it is legal by 2020 and I expect the entire world to allow such driverless vehicles everywhere by 2025.

electric-car-insider.com

Henrik, we usually only hear about the AP failures when the consequences are truly tragic, or if someone is running a dash cam and posts the video. But there are others, and will be more.
Every edge case that can't be handled by AP, which also has an inattentive driver, will result in a collision. I predict that every edge case will eventually be discovered the hard way.

When a pilot wants to fly a technically sophisticated airplane, they have to get an extra "rating" or at very least a "checkout". This is in addition to the initial pilot certification. Both involve training and testing (different formats, rigor).

Why are we allowing drivers to drive technically sophisticated cars without demonstrating understanfing of the systems' limitations and functional competence?

Account Deleted

We are allowing it because if we demanded that we did not do anything until we were sure no accidents would happen we would end up in a world where nothing gets done and where making progress would be impossible.

We as humans invent most things out of accident and by systematic trial and error processes. We are not so clever that we can figure it all out in the lab and then launch an error free product. Tesla is doing it the right way IMO. By applying the AP in the real world and using OTA software updates they can advance and perfect this technology much faster and that will save million of lives compared to a few hundred that may die because of technical glitches and not yet covered fringe cases. As you know ECI 1.3 million lives are lost every year globally to traffic accidents so we really need to develop AP tech as fast as possible.

We also need autonomous cars to make otherwise prohibitively expensive BEVs economical so that they can take over from gassers and save the planet from a greenhouse apocalypse. BEVs only get an economic advantage over gassers if you drive them a lot per year so driverless BEV taxis that can do over 100k miles per year will be much less costly than driverless gasser taxis that does over 100k per year. Battery tech will not get much cheaper after Tesla’s giga factory is up and running and it is not enough for BEVs to win over gassers. You need driverless cars that drives over 50k miles per year for that to happen.

electric-car-insider.com

Henrik> if we demanded that we did not do anything until we were sure no accidents would happen...

Oh boy. That's a whopper of a straw man. No one has suggested we "not do anything". Quite the opposite. I don't know if you sidestepped my point intentionally, but I'll give you a chance for a do-over if you'd like.

The comments to this entry are closed.