Exactly. There's no way the AP2 car would be so rudimentary like it is now if their video is real. They are definitely starting from near-scratch with the AP2 hardware currently.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
This is my observation of how the code works.
Typical case:
AP tracks a moving object in front of it. If object slows down, AP slows down. If object stops, AP stops. If object accelerates, AP accelerates. If object never moves, AP doesn't move.
Danger case:
AP has no moving object to track. It will collide at full speed with immobile object with no attempts at detection or mitigation.
I've tested the danger case by seeing if AP will slow down / stop for a vehicle that is at a Red light and also initially outside of Tesla vision. It doesn't. Going 50MPH catches up to a 0MPH vehicle really fast. Slammed on the brakes. I am so concerned about this aspect that I would highly recommend against wider release of the firmware unless Tesla can address this.
You do realize that Tesla has multiple branches to the code, right? The software version you're running on isn't the same as what they are currently doing QA testing on, which is also different from the FSD code they're developing right now. The video was probably the latest and greatest version of the software, which still has a long way to go before it's ready for release.
So they are giving the public the oldest and worst builds? Or are they giving the public the alpha builds? Or what... most STABLE builds? Whatever the case is one can only image how unstable the builds are in the "latest and greatest".
Instead of that, think of multiple branches. There is a EAP branch, a TACC branch. These branches have various things turned off. It isn't about the stability of a branch.
ZeroDarkSilver
HEPA cleaning your atmosphere, you're welcome
Keep in mind that at no point in that video does the car exceed 35mph.
On the issue of braking for stationary objects:
You might want to read first the 8.0 software release announcement and press conference. We were all shaken and saddened when Jason Brown was killed when ran into the side of a truck in his Tesla running firmware 7,1. The truck's color provided insufficient contrast with the sky background and essentially the camera, which was the primary arbiter of objects in front of the car, didn't "see" it. RADAR saw it, but the car didn't stop because there was no confirmation from the camera.
In response to this condition, Tesla increased the role of the RADAR so that the car can now stop based on a RADAR signal alone. The problem for Tesla is that an overhead metal sign appears the same as an object on the road to RADAR when the car is cresting a hill. It would be very dangerous if the car slammed on the brakes every time one went up a hill where there is an overhead sign. So Tesla is using the feet to map the overhead signs in the world. If cars pass under these signs repeatedly, then the Tesla database can be sure that it's a sign and not object in the road 99+% of the time.
So right now, we are in between the old camera system and the fully implemented new RADAR system for AP 1.0 cars (and I presume for AP 2.0).
I don't want to leave the impression that AP 1.0 cars regularly ignore stationary objects in the road because they do not. Just about all the time if the camera and the RADAR can see the object, the car will stop. Sometimes the stop is a little closer to the object than our intuition says is safe, but usually the car will stop. In 6,000 miles of driving, much with AP, there have only been two instances where I decided to intervene and brake the car, Tesla Autopilot, whether 1.0 or 2.0 is designed and specified for a human to monitor and to take over when the car fails to handle a situation correctly.
One reasonably hopes that AP 1.0 will improve once sign mapping is completed and the software supports that data, and EAP based on AP 2.,0 hardware will far outstrip the original Autopilot.
Also check out this video of a Tesla stopping for pedestrians.
You might want to read first the 8.0 software release announcement and press conference. We were all shaken and saddened when Jason Brown was killed when ran into the side of a truck in his Tesla running firmware 7,1. The truck's color provided insufficient contrast with the sky background and essentially the camera, which was the primary arbiter of objects in front of the car, didn't "see" it. RADAR saw it, but the car didn't stop because there was no confirmation from the camera.
In response to this condition, Tesla increased the role of the RADAR so that the car can now stop based on a RADAR signal alone. The problem for Tesla is that an overhead metal sign appears the same as an object on the road to RADAR when the car is cresting a hill. It would be very dangerous if the car slammed on the brakes every time one went up a hill where there is an overhead sign. So Tesla is using the feet to map the overhead signs in the world. If cars pass under these signs repeatedly, then the Tesla database can be sure that it's a sign and not object in the road 99+% of the time.
So right now, we are in between the old camera system and the fully implemented new RADAR system for AP 1.0 cars (and I presume for AP 2.0).
I don't want to leave the impression that AP 1.0 cars regularly ignore stationary objects in the road because they do not. Just about all the time if the camera and the RADAR can see the object, the car will stop. Sometimes the stop is a little closer to the object than our intuition says is safe, but usually the car will stop. In 6,000 miles of driving, much with AP, there have only been two instances where I decided to intervene and brake the car, Tesla Autopilot, whether 1.0 or 2.0 is designed and specified for a human to monitor and to take over when the car fails to handle a situation correctly.
One reasonably hopes that AP 1.0 will improve once sign mapping is completed and the software supports that data, and EAP based on AP 2.,0 hardware will far outstrip the original Autopilot.
Also check out this video of a Tesla stopping for pedestrians.
Last edited:
There are different kinds of city streets. I use TACC on some city streets, largely determined by the absence pedestrian traffic. And when I use it, TACC works quite well on those streets.
I will say that the longer I use AutoPilot, the more I use it in accordance with the Owners Manual. There are things that work well, and things that are safe. The two classes are not always the same.
krazineurons
krazineurons
What happens when I have enabled TACC, driving under 35MPH and eventual run into stop go traffic. Would it break then?
My take from everything I've been reading from MXWing is that it will not stop. It will only stop if you engaged TACC while moving and with a car in front of you at the time that you engaged TACC. When that car stops, your call will stop. If you however engage TACC at at time where there are no cars in front of you, say the next car in front is 1/4 mile down the road and you come up on this car, which is now stopped due to congested traffic, the Tesla will not stop since it had no car to originally track when you engaged TACC.
The question I have is what made you think TACC could stop for a car that isn't moving. TACC is designed to follow cars moving, and isn't designed to stop for an obstruction or debris on the road. It can't detect a stalled car.
There is anything magical about AP2 in it's current state that makes your TACC any different than the TACC in AP1. The TACC in AP1 has led to a few people crashing into stalled cars because for some reason they got the impression that it would stop.
TACC on an AP1/AP2 car isn't that much different than the radar based adaptive cruise control in a lot of cars, and these can't detect stopped objects either.
Does the manual leave out the fact that TACC can't detected/react to stopped cars?
The biggest danger right now with how they're releasing it is the reality of it isn't matching peoples expectations.
At some point in some future version it will stop, but that isn't what this version has.
Last edited:
Tjhappel
Member
I have been driving at least 200 mikes a day, 85% on Tacc on the freeways, it's super super annoying, from san clemente to venice I had ~10 false positives all the bike I was in the carpool lane. Interesting to note all of them happened when Tesla vision was "not" following someone else. Today from san clemente to riverside it was about the same, last night I learned about the "report bug" ability so I tried to do it every time I could. It does seem we have a very very very far way to go and I'd be lying if I wasn't second guessing my full ap purchase instead of waiting to see if they could ever even do it, the video and ap 1 did imply they were much further ahead than what I am currently seeing. With that being said I'm happy to be a beta tester for them, I love the company and how aggressive they are so I'll keep being beta tester for now and hope it gets real good relatively quickly.
Thanks for good posts everyone, helpful. On a side not tacc seems pretty damn good except for the overhead false positives. Not too jerky compared to competition.
Thanks for good posts everyone, helpful. On a side not tacc seems pretty damn good except for the overhead false positives. Not too jerky compared to competition.
When you say 10 false positives do you mean the TACC started slowing for a car that wasn't there? Or 10 false positives of the forward collision warning (just a warning that doesn't slow down the car).
I would email them instead of reporting the bug. The email address on the release notes is SPECIFICALLY for reporting these issues. Where the other one has to go through layers to get to the right department.
S100D
Member
LastGas, if you try the following scenario I think you will find out that you will destroy your car.
Turn on the autopilot on AP1 and find a car at a stoplight that is about 300 feet away. Your car cannot see the car stopped at the stoplight since it is too far away. Since you were not following the car when it stopped at the stoplight, you Tesla will not stop and it will run into the back of the car at the stop light. Remember that AP1 does not have automatic emergency braking. The best that it might do is reduce the speed by 25 mph before it hits the car. Worst case is that it will not slow and and will run into the car.
I made the mistake to think that my $100,000 Tesla had emergency braking. The first day that I had my S85D, I had the autopilot on and almost ran into the back of a car that was stopped at a stoplight. I slammed on the brakes and was able to stop a few feet from destroying my car.
Andyw2100
Well-Known Member
My email and response from Tesla
My Email:
Dear Tesla,
I wanted to write my feedback to you after two days of driving with firmware 2.50.185 with the hopes the information will contribute to the success of the Tesla vision.
False Positive Detection on Forward Collision Detect and Sudden Breaking
1.) Headed towards San Diego super charger from the Midway museum on Sunday Jan 1st.
2.) Forward Collision Detect set to early.
3.) TACC enabled at 70mph.
4.) On 3 occasions I hear beeping and see a red car show up on my dashboard. Car deaccelerstes sharply each time. I hit brake the moment I can react. Nothing is in front of me.
5.) I turn off forward collision assist for the remainder of the trip home and try to evaluate TACC performance independent of collision assist. I never saw the issue again with using TACC only.
My hypothesis is there might of been an overpass or cars in front but "above" me which set something off. I know GPS systems can be confused by objects/road on the Z-axis of travel.
TACC and Immobile Cars in front of it.
I turned on TACC on a road that has a 50mph speed limit and it has lots of red lights.
As TACC was coming up to a stopped car it did not do any slowing down.
I was forced to brake hard. As a human driver I would have seen the car far enough in advance to let off the "gas", And then come to a full stop with very light braking. I do not believe TACC had any reaction to the immobile car.
However, when the light turned green, and we both hit 25 - TACC behaved how I would expect. It maintained sufficient following distance, accelerated and decelerated as needed and could handle full stops and starts from full stops.
Tesla's Response:
Thank you for reaching out with your feedback. Please let us know if your vehicle continues to detect objects which do not appear to be real and we would be happy to have service double check your sensors.
As for the example that you mentioned where you came up upon a stopped vehicle, TACC is not optimized for detecting them at this point. I have pasted the description from the manual which covers this in particular but I do recommend reading the correct operation of the driver assist functions in the manual as well.
"Warning: Traffic-Aware Cruise Control cannot detect all objects and may not brake/decelerate for stationary vehicles, especially in situations when you are driving over 50 mph (80 km/h) and a vehicle you are following moves out of your driving path and a stationary vehicle or object is in front of you instead. Always pay attention to the road ahead and stay prepared to take immediate corrective action. Depending on Traffic-Aware Cruise Control to avoid a collision can result in serious injury or death. In addition, Traffic-Aware Cruise Control may react to vehicles or objects that either do not exist or are not in the lane of travel, causing Model X to slow down unnecessarily or inappropriately."
My Email:
Dear Tesla,
I wanted to write my feedback to you after two days of driving with firmware 2.50.185 with the hopes the information will contribute to the success of the Tesla vision.
False Positive Detection on Forward Collision Detect and Sudden Breaking
1.) Headed towards San Diego super charger from the Midway museum on Sunday Jan 1st.
2.) Forward Collision Detect set to early.
3.) TACC enabled at 70mph.
4.) On 3 occasions I hear beeping and see a red car show up on my dashboard. Car deaccelerstes sharply each time. I hit brake the moment I can react. Nothing is in front of me.
5.) I turn off forward collision assist for the remainder of the trip home and try to evaluate TACC performance independent of collision assist. I never saw the issue again with using TACC only.
My hypothesis is there might of been an overpass or cars in front but "above" me which set something off. I know GPS systems can be confused by objects/road on the Z-axis of travel.
TACC and Immobile Cars in front of it.
I turned on TACC on a road that has a 50mph speed limit and it has lots of red lights.
As TACC was coming up to a stopped car it did not do any slowing down.
I was forced to brake hard. As a human driver I would have seen the car far enough in advance to let off the "gas", And then come to a full stop with very light braking. I do not believe TACC had any reaction to the immobile car.
However, when the light turned green, and we both hit 25 - TACC behaved how I would expect. It maintained sufficient following distance, accelerated and decelerated as needed and could handle full stops and starts from full stops.
Tesla's Response:
Thank you for reaching out with your feedback. Please let us know if your vehicle continues to detect objects which do not appear to be real and we would be happy to have service double check your sensors.
As for the example that you mentioned where you came up upon a stopped vehicle, TACC is not optimized for detecting them at this point. I have pasted the description from the manual which covers this in particular but I do recommend reading the correct operation of the driver assist functions in the manual as well.
"Warning: Traffic-Aware Cruise Control cannot detect all objects and may not brake/decelerate for stationary vehicles, especially in situations when you are driving over 50 mph (80 km/h) and a vehicle you are following moves out of your driving path and a stationary vehicle or object is in front of you instead. Always pay attention to the road ahead and stay prepared to take immediate corrective action. Depending on Traffic-Aware Cruise Control to avoid a collision can result in serious injury or death. In addition, Traffic-Aware Cruise Control may react to vehicles or objects that either do not exist or are not in the lane of travel, causing Model X to slow down unnecessarily or inappropriately."
HX_Guy explained my experience very well. TACC works only if it has an object to track against.
Please do report this to Tesla. The more feedback the better.
krazineurons
krazineurons
So the core problem of Tesla not breaking into stationary objects with TACC is that it can't deterministically identity if there a stationary vehicle in front of it.
Am guessing that is where having the system learn to identify cars from the tri focal cameras and Radar data processing comes in and eventually training the model to learn identifying the cars would help and perhaps that is why few million miles are needed before it happens.
But how did the AP2 demo video stop for stopped vehicles in front of it, clearly the code and data model exists, then why not put that in beta with this update.
Alternatively they could provide a visual cue notifying the driver that the car is not going to break and one should apply the breaks as an object is detected in front. Or is that FCW?
Am guessing that is where having the system learn to identify cars from the tri focal cameras and Radar data processing comes in and eventually training the model to learn identifying the cars would help and perhaps that is why few million miles are needed before it happens.
But how did the AP2 demo video stop for stopped vehicles in front of it, clearly the code and data model exists, then why not put that in beta with this update.
Alternatively they could provide a visual cue notifying the driver that the car is not going to break and one should apply the breaks as an object is detected in front. Or is that FCW?
Chaz
Big Kahuna
The current version of TACC, as mentioned by others and the owner's manual, does not brake your Tesla completely for stationary objects, particularly objects it did not track as coming to a stop. Although AP works great in stop-and-go traffic (with a careful understanding of it's limitations, in my opinion), it was not designed for non-highway/freeway situations....yet.
The car in the self-driving demo video was demonstrating one particular version of the code, probably with other features/capabilities turned off, and is most likely undergoing extensive testing and de-confliction with other systems in the vehicle.
If I had to guess...
Probably because it's not fully tested. There is one thing worse than not stopping for stationary vehicles, and that is randomly stopping for stationary vehicles. The system has to work flawlessly and predictably. If it stops for stationary objects most of the time, then drivers become complacent and it's a bigger problem then.
