Model d not likely hardware sufficient to be driverless

Why do you need a 360 degree long range radar? The google car uses just one long range lidar, but the Model S instead has forward radar, 360 degree medium range ultrasonics, and video cameras. Why do you need more?

You're partially right. There is more hardware (probably) needed, but it doesn't have to be $75k
Why does it have to be laser? I don't have laser beams in my eyes, and I drive just fine.
The radar and camera combination used by Tesla is much cheaper than those lidar units, and just as effective (arguably more so actually). The problem right now is that it's only forward facing, The vehicle probably has sufficient rear coverage already as well, but it's missing cross-traffic, so it probably really needs 2 more cameras, or 2 more radar units (or both) to work fully driverless. (I don't know that the current setup could properly (and safely) handle a 4 way intersection, especially one without traffic lights) The ultrasonics, while great, aren't really long range enough for this use.

That said, the current setup is quite likely enough to get to the "90% driverless" that musk keeps talking about.

ran1354 said:

I noticed that it has a forward radar and other sensors. However
to be driverless a 360 degree laser would be required,like the Google
driverless cars. A laser of that caliber alone would be 75k$. So
the d is a step in the right direction but a future hardware upgrade will
be required .

Click to expand...

There is no claim or intent for D to be Driverless

The rear coverage as described in the autopilot presentation isn't sufficient for merging into lanes with fast moving traffic. On the autobahn, careless merging is a recipe for being nailed by a car going 50 mph faster than you are.

It's *possible* the reverse camera is sufficient, but probably you'd need a rear facing radar, at least in adverse weather conditions.

Yggdrasill said:

The rear coverage as described in the autopilot presentation isn't sufficient for merging into lanes with fast moving traffic. On the autobahn, careless merging is a recipe for being nailed by a car going 50 mph faster than you are.

It's *possible* the reverse camera is sufficient, but probably you'd need a rear facing radar, at least in adverse weather conditions.

Click to expand...

Radar isn't actually necessary either front or rear, but it is nice to have. Remember, you don't have radar in your eyes now, and you manage just fine. As for if the rear camera is enough... that's a tough one, it's designed for straight behind you, not off to the side (and is really geared for short range)

I agree more hardware is necessary for full self driving. I've been waiting for the auto-pilot announcement for months now (c'mon, we all knew it was coming!) and I'm both excited by what they've done, and disappointed they didn't do more. I had thought that they'd have all the necessary hardware for "self driving", and the rest would be software, but that's just not the case. That said, they've leapfrogged the competition on driver assist features (or at least will if they manage to actually get the software updates out the door before the competition gets any better) and that's quite a feat right there, so it's hard to quibble over hardware that they probably wouldn't be able to fully use for a couple years anyway.

I'll say hardware wise that they've met my minimum requirements, but software wise they've exceeded my expectations (assuming the promised rollout).

Elon said at the event that the cars have 360 degree sonar. The car can detect the density of objects.
You can see an image and description on the Tesla Motors Model S "Order" page.

v12 to 12v said:

Elon said at the event that the cars have 360 degree sonar. The car can detect the density of objects.
You can see an image and description on the Tesla Motors Model S "Order" page.

Click to expand...

The problem is that the 360 degree ultrasonics only have a range of 18ft, that's plenty to avoid changing lanes in to a barrier, or backing over a child in a parking lot, but not sufficient to deal with high speed traffic at an intersection.

The goal appears to be auto pilot capability after merging into traffic and until exiting... essentially the long boring stretches of driving... which there appears to be sufficient hardware for.

It's driver assist not fully autonomous driving. I'm sure human action is needed to make any large changes, such as moving to merge at a large interchange or something and other hazardous situations.

Plus, they've made this North America only so far I believe... so no autobahn issues here.