Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

How can Autopilot be applauded? Any suggestions?

This site may earn commission on affiliate links.
#41
seattlite2004 said:
Maybe....maybe not....and if there was a car in the right lane? Then what?
Click to expand...
The car didn't go into the right lane. The driver did. The car only notified the driver and shifted to the right side of the lane it was already in. Read the description that grommet posted.

The car has sensors so it would know if there were any cars in the right lane and use that information to determine if it was safe to go to the right lane or not.
 
#43
Garlan Garner said:
The question remains...

What if there was a car in the right lane? Then what?

It doesn't matter if the Human or Autopilot drove into the right lane. The question remains.
Click to expand...
Well the car didn't change lanes. If there was a car in the right lane it still wouldn't have changed lanes.

"Our Model 3 Autopilot when it sensed an accident was imminent, sent out an audible warning and steered us to the right within the lane when we then took over for Autopilot the moment I honked our horn, simultaneously looked over rear view mirror in right adjacent lane to make sure no vehicle was there before moving over to right adjacent lane to continue to avoid collision and powered out getting back into the same lane. The Autopilot disengage audible sound was washed out from the horn sound I believe because of our low quality dashcam mic."
 
#49
chronopc said:
Not sure. I believe it all depends on the situation. I'm not going to test it out though. :D
Click to expand...
There ya go. There is the answer most everyone has. And its the correct answer.

"I don't know".

No one knows what autopilot "would have done"
No one know what "the human" would have done.

How can anyone say which driver would have been best? You can't....however what we DO know is that.


Autopilot "recognized" the danger and alerted the driver. That's fantastic in my eyes.
 
#50
Garlan Garner said:
There ya go. There is the answer most everyone has. And its the correct answer.

"I don't know".

No one knows what autopilot "would have done"
No one know what "the human" would have done.

How can anyone say which driver would have been best? You can't....however what we DO know is that.


Autopilot "recognized" the danger and alerted the driver. That's fantastic in my eyes.
Click to expand...
Tesla should know what EAP would have done. It should be part of the programming. All of what EAP does is part of a series of it this, then that.

Humans on the other hand is all random and that's why accidents happen everyday.
 
#51
Garlan Garner said:
There ya go. There is the answer most everyone has. And its the correct answer.

"I don't know".

No one knows what autopilot "would have done"
No one know what "the human" would have done.

How can anyone say which driver would have been best? You can't....however what we DO know is that.


Autopilot "recognized" the danger and alerted the driver. That's fantastic in my eyes.
Click to expand...
Wonder if other vehicle's AEB would have also beeped...maybe braked...but of course they would not have swerved.
 
#53
chronopc said:
Tesla should know what EAP would have done. It should be part of the programming. All of what EAP does is part of a series of it this, then that.

Humans on the other hand is all random and that's why accidents happen everyday.
Click to expand...
I don't think so.

When Autopilot grows up into FSD.....things aren't going to be as black and white as programing.

There are a ton of "IF....Then" interwoven statements built in. Just like there are in playing the game of chess. The computer doesn't have a billion situations built in and then plays the game of chess. Its a complex web of "if......then" statements that it uses what move to make. billions of them each time its the computers chance to move.

Once the chess computer finishes its move....it starts ALL over calculating its next move based on "if...then" and positional advantages it can force.

Autopilot is currently recognizing its environment and making limited changes to its position in traffic on a second by second basis and often its decisions are saving lives.
How do we applaud that?
 
#55
Garlan Garner said:
I don't think so.

When Autopilot grows up into FSD.....things aren't going to be as black and white as programing.

There are a ton of "IF....Then" interwoven statements built in. Just like there are in playing the game of chess. The computer doesn't have a billion situations built in and then plays the game of chess. Its a complex web of "if......then" statements that it uses what move to make. billions of them each time its the computers chance to move.

Once the chess computer finishes its move....it starts ALL over calculating its next move based on "if...then" and positional advantages it can force.

Autopilot is currently recognizing its environment and making limited changes to its position in traffic on a second by second basis and often its decisions are saving lives.
How do we applaud that?
Click to expand...
It is as simple as programming. Why do you think data is so important? It provides company like Tesla with situations they need to deal with and they do this by changing the algorithm. I"m really close to Silicon Valley. Everything here is programming.
 
#56
chronopc said:
It is as simple as programming. Why do you think data is so important? It provides company like Tesla with situations they need to deal with and they do this by changing the algorithm.
Click to expand...
Wow....

Programming isn't simple. NOT AT ALL.

Anyone can program a car to go straight down a street when there is a guarantee that there will be no issues. . Situations like this video is very difficult because its situational.

If the vehicle that swerved over from the left lane were 3 feet further ahead....then.....
If the vehicle that swerved over from the left lane were 6 feet further back.....then....
If the vehicle that swerved over from the left was traveling 20 MPH faster......then....
If the vehicle that swerved over from the left lane was a truck.....then...
If my car was already in the rightmost lane...then...…
If my car was exiting onto an off ramp when this happened.....then.....
If the vehicle that swerved over from the left lane was a motorcycle.....then....
If both a car from the right and left did the same thing at the same time.....then.....
If....there was a car already in the right lane...then.....
If....my car was stopped at a red light and this happened......then...
If...the vehicle that swerved over was a bicycle and I was only traveling 5MPH.....
If the same scenario happened on a 2 lane highway.....then...…
If...my cars brakes failed….then....
If...the person in the car that swerved was having a stroke a medical emergency....then....
If....

easy? I don't think so.

SO many times people think navigating a 2 ton vehicle is easy because they haven't been faced with major accidents. Although not having accidents is wonderful it often lends itself to complacency.

Computers don't learn...and therefore can't become complacent. They are just as alert at the 1st second than they are at the millionth.

How can we push and applaud what Tesla autopilot is doing on a daily basis?
 
#57
Garlan Garner said:
Wow....

Programming isn't simple. NOT AT ALL.

Anyone can program a car to go straight down a street when there is a guarantee that there will be no issues. . Situations like this video is very difficult because its situational.

If the vehicle that swerved over from the left lane were 3 feet further ahead....then.....
If the vehicle that swerved over from the left lane were 6 feet further back.....then....
If the vehicle that swerved over from the left was traveling 20 MPH faster......then....
If the vehicle that swerved over from the left lane was a truck.....then...
If my car was already in the rightmost lane...then...…
If my car was exiting onto an off ramp when this happened.....then.....
If the vehicle that swerved over from the left lane was a motorcycle.....then....
If both a car from the right and left did the same thing at the same time.....then.....
If....there was a car already in the left right lane...then.....
If....my car was stopped at a red light and this happened......then...
If...the vehicle that swerved over was a bicycle and I was only traveling 5MPH.....
If the same scenario happened on a 2 lane highway.....then...…
If...
If...
If....

easy? I don't think so.
Click to expand...
I didn't say that programming is simple. I'm saying that programming or software is what EAP is. So yes, Tesla would know what EAP would have done in that situation because they wrote the programming for EAP. All of EAP's actions are recorded on a log. So Tesla can pull up the information and see what EAP saw and why it decided to do what it did and they can adjust that behavior if they want to.

Like you mentioned with computers playing Chess, programming is what tells it to reassess it's best decision based on changes in the situation. So if you can program a computer can play Chess, why can't you program a computer to make driving decisions?
 
#58
chronopc said:
I didn't say that programming is simple. I'm saying that programming or software is what EAP is. So yes, Tesla would know what EAP would have done in that situation because they wrote the programming for EAP. All of EAP's actions are recorded on a log. So Tesla can pull up the information and see what EAP saw and why it decided to do what it did and they can adjust that behavior if they want to.
Click to expand...

So how do we push the success of Autopilot for what it has already successfully accomplished in the real world? Tesla doesn't have to evaluate the successes. Its doing it. There is nothing to change.

Lets applaud when Autopilot doesn't crash into a barrier as it has been falsely touted that it has in article after article. Tesla has a hard line to walk when they KNOW from data that a person didn't follow instructions and grab the wheel for whatever reason. Tesla can't risk being insensitive to its customers even when they are known to be at fault in a tragedy.

How can WE help Tesla represent the successes of Autopilot?
 
#60
Garlan Garner said:
So how do we push the success of Autopilot for what it has already successfully accomplished in the real world? Tesla doesn't have to evaluate the successes. Its doing it. There is nothing to change.

Lets applaud when Autopilot doesn't crash into a barrier as it has been falsely touted that it has in article after article. Tesla has a hard line to walk when they KNOW from data that a person didn't follow instructions and grab the wheel for whatever reason. Tesla can't risk being insensitive to its customers even when they are known to be at fault in a tragedy.

How can WE help Tesla represent the successes of Autopilot?
Click to expand...
Tesla is constantly changing EAP. I've only had my Model 3 for over a month and I've already received an update to EAP and I look forward to receiving more in the future. EAP should continue to change as Tesla gets more data. For example that one crash that killed the person in the Model X. Tesla used the information it got from that incident and tweaked Autopilot's programming to ensure a situation doesn't occur.

I personally love EAP but it's not perfect right now, but it can and will improve. Tesla will do that by getting more data and changing the programming of EAP.

I paid $5k for EAP if that's not touting Tesla's EAP, I don't know what is. Lol.
 
You must log in or register to reply here.

Similar threads

J
NHTSA make roads across the world less safe
Replies
51
Views
2K
AI, Autopilot, & Autonomous/FSD
DoctorWTF
D
B
Plate Changeout - How?
Replies
35
Views
3K
The UK and Ireland
debrooker
D
R
Tesla Service Screwed up
Replies
1
Views
504
Model X
scottf200
scottf200
L
M3 Hit and Run, how bad is it?
Replies
21
Views
2K
Model 3
nate704
N
E
  • Article
Tesla kids science book + Our Model 4 prediction!
Replies
0
Views
243
Tesla, Inc.
EV_Twins
E
Top
Back
Blog
New
Forum list
Marketplace
Menu