Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Tesla has a self-driving strategy other companies abandoned years ago

arstechnica.com

Post reply on HN
Print view View original

31–40 of 222 posts

Re: Tesla has a self-driving strategy other companies abandoned years ago

#31

Post by kec »

Robotbeat wrote: post #21

Interestingly, for air travel, there's another strategy: Rely entirely on GPS and commanding from air traffic control (with systems and structure in place to minimize damage in case those systems fail or are insufficient). That's what Zipline uses for their fully operational medical delivery system in Rwanda. The drones are completely blind, directly, but they travel in pre-defined flight paths and communicate with o…

If a plane can only get a GPS fix to within a 15m radius that's no big deal, whereas if a car has a 15m fix it's running down pedestrians on the sidewalk. Planes also don't have a habit of flying in tunnels or urban canyons where obtaining a gps fix is even more problematic.

Re: Tesla has a self-driving strategy other companies abandoned years ago

#32

Post by peeters »

I've always found that half the issue is with the name: Autopilot. On an airplane, autopilot means you can take your focus off of both the physical and mental act of moving the plane through the air, and if it needs you to intercede there is pretty much no emergency in which you wouldn't have a few seconds to react.

In a car, Tesla's Autopilot simply doesn't give you that. If you are reading and your autopilot exits, you could be dead before your focus can return to the road. Planes fly in wide open spaces far from hazards. Cars drive in busy, tight spaces where a fraction of a second lapse in control can be fatal.

I get that Tesla's implementation is analogous to what you find on a jet. But the environments make them very different. So start by dropping the name. It's not auto pilot, it's handsfree driving.

Re: Tesla has a self-driving strategy other companies abandoned years ago

#33

Post by 0815test »

peeters wrote: post #32

I've always found that half the issue is with the name: Autopilot. On an airplane, autopilot means you can take your focus off of both the physical and mental act of moving the plane through the air, and if it needs you to intercede there is pretty much no emergency in which you wouldn't have a few seconds to react. In a car, Tesla's Autopilot simply doesn't give you that. If you are reading and your autopilot exits,…

> I've always found that half the issue is with the name: Autopilot. ...In a car, Tesla's system simply doesn't give you that [taking your focus off]

Yes, exactly. It's a level-2 ADAS - it's about helping you deliver correct inputs to the car, not about autonomous driving in any real sense - regardless of how advanced it might be in a technical sense, or how well-behaving in a best-case scenario! More like the fly-by-wire system in an aircraft than the "autopilot".

And I know that the users are supposed to be aware of this, but Tesla's marketing on this issue is far from consistent, in all sorts of ways. They're doing their users a pretty big disservice.

Re: Tesla has a self-driving strategy other companies abandoned years ago

#34

Post by Robotbeat »

kec wrote: post #31
Robotbeat wrote: post #21

Interestingly, for air travel, there's another strategy: Rely entirely on GPS and commanding from air traffic control (with systems and structure in place to minimize damage in case those systems fail or are insufficient). That's what Zipline uses for their fully operational medical delivery system in Rwanda. The drones are completely blind, directly, but they travel in pre-defined flight paths and communicate with o…

If a plane can only get a GPS fix to within a 15m radius that's no big deal, whereas if a car has a 15m fix it's running down pedestrians on the sidewalk. Planes also don't have a habit of flying in tunnels or urban canyons where obtaining a gps fix is even more problematic.

Indeed. But this is an addressable problem. The new GPS satellites (GPS III) have just started launching (on a SpaceX rocket in December 2018), and they should triple the accuracy: https://www.geospatialworld.net/blogs/what-is-gps-iii-and-ho...

But even before that's complete: High precision GPS (i.e. with compensation for atmospheric effects) can get better than 10cm accuracy by tracking the carrier wave: https://en.wikipedia.org/wiki/Real-time_kinematic

Single-digit-centimeter-accuracy is possible near GPS signal compensating base stations (~20km). This level of precision is regularly used by surveyors.

There are several options for improving geolocation beyond inside-out (i.e. vision or lidar) tracking.

Re: Tesla has a self-driving strategy other companies abandoned years ago

#35

Post by _ea1k »

peeters wrote: post #32

I've always found that half the issue is with the name: Autopilot. On an airplane, autopilot means you can take your focus off of both the physical and mental act of moving the plane through the air, and if it needs you to intercede there is pretty much no emergency in which you wouldn't have a few seconds to react. In a car, Tesla's Autopilot simply doesn't give you that. If you are reading and your autopilot exits,…

Autopilot in planes can be used, and often is used, in cases of VFR flight in which see-and-avoid is a requirement. Constant surveillance of the aircraft's surroundings is important. When it isn't done, midairs happen.

Re: Tesla has a self-driving strategy other companies abandoned years ago

#36

Post by 0815test »

_ea1k wrote: post #35
peeters wrote: post #32

I've always found that half the issue is with the name: Autopilot. On an airplane, autopilot means you can take your focus off of both the physical and mental act of moving the plane through the air, and if it needs you to intercede there is pretty much no emergency in which you wouldn't have a few seconds to react. In a car, Tesla's Autopilot simply doesn't give you that. If you are reading and your autopilot exits,…

Autopilot in planes can be used, and often is used, in cases of VFR flight in which see-and-avoid is a requirement. Constant surveillance of the aircraft's surroundings is important. When it isn't done, midairs happen.

That's a special case, however. See-and-avoid except in specialized scenarios is what the Level-3 designation of driving-assistance systems is about, something that Tesla has not achieved so far.

Re: Tesla has a self-driving strategy other companies abandoned years ago

#37

Post by lutorm »

kec wrote: post #31
Robotbeat wrote: post #21

Interestingly, for air travel, there's another strategy: Rely entirely on GPS and commanding from air traffic control (with systems and structure in place to minimize damage in case those systems fail or are insufficient). That's what Zipline uses for their fully operational medical delivery system in Rwanda. The drones are completely blind, directly, but they travel in pre-defined flight paths and communicate with o…

If a plane can only get a GPS fix to within a 15m radius that's no big deal, whereas if a car has a 15m fix it's running down pedestrians on the sidewalk. Planes also don't have a habit of flying in tunnels or urban canyons where obtaining a gps fix is even more problematic.

Planes already have a problem with GPS outages or military jamming exercises. Imagine if all cars within a 100 mile radius all of a sudden would stop working.

Re: Tesla has a self-driving strategy other companies abandoned years ago

#38

Post by Someone1234 »

martythemaniak wrote: post #30

There's a few things to keep in mind when discussing self-driving tech. 1) Self driving tech doesn't exist today. It simply does not. There's lots of people working on it and they are using different strategies, but it is not clear how long it will take, which strategy is technically superior, which strategy is more economical, etc etc. There's a lot of very strong opinions floating around (LIDAR is essential! No, vi…

> They're in the business in shipping real cars to real people and there was never any possibility of mixing that with LIDAR.

Can you explore that a little? We have consumer products that contain LIDAR already (e.g. Neato Vacuums). The sensors no longer cost thousands of dollars, and could supplement other types of inputs to provide a broader picture of the road.

Re: Tesla has a self-driving strategy other companies abandoned years ago

#39

Post by razorbladeknife »

ForHackernews wrote: post #18
cbames89 wrote: post #5

This is a (perhaps short-term) failure of first principles reasoning. Elon's known to favor thinking about things from first principles, and there's no theoretical reason that vision can't work. However, technical limitations might take a while to overcome.

Vision alone fails pretty frequently even with a human brain and 700 million years of evolution behind it. Count me as a skeptic that cameras + computer will produce a safe self-driving system.

We can't measure that evolutionary advantage.

What we can measure is fatalities

Re: Tesla has a self-driving strategy other companies abandoned years ago

#40

Post by pbreit »

martythemaniak wrote: post #30

There's a few things to keep in mind when discussing self-driving tech. 1) Self driving tech doesn't exist today. It simply does not. There's lots of people working on it and they are using different strategies, but it is not clear how long it will take, which strategy is technically superior, which strategy is more economical, etc etc. There's a lot of very strong opinions floating around (LIDAR is essential! No, vi…

I don't have a huge problem with the supposed goalpost movement. I am wondering though if drivers need to keep a fairly close eye on things what the true benefit of self-driving is in near future.
Post reply on HN