Because they are real assist systems that one drive together with the car, not wannabe level 4 like autopilot and FSD that still is only level 2 capable.Maybe they're all programmed to comply with EU regulations.
You can install our site as a web app on your iOS device by utilizing the Add to Home Screen feature in Safari. Please see this thread for more details on this.
Note: This feature may not be available in some browsers.
Because they are real assist systems that one drive together with the car, not wannabe level 4 like autopilot and FSD that still is only level 2 capable.Maybe they're all programmed to comply with EU regulations.
Mass market lidars have improved a lot from your example. But that's beside the point. Lidar can help even without identifying the lane. Consider the example below. I'm pretty convinced Tesla's vision NN mistook the vertical glare off the headlights for two lane lines and steered in between them. Lidar would have identified the large, solid object in this misidentified "lane". Lidar localizing on an HD map would have told the supervisory code that the misidentified lane was in the wrong place. This is all very valuable information, and separate from lidar itself identifying the actual lane lines.While some LIDAR sensors have enough resolution and provide enough data (using reflectivity) to detect lane lines, the kind of consumer LIDAR that are being used in massed produced passenger cars today (like the Valeo Scala used by Audi and no doubt would be the type Ford would use if they opted for it) simply does not have the resolution to reliably detect lane lines even if they were perfect and on a clear day with clean roads. You can see some sample outputs in Valeo's presentations:
https://www.mathworks.com/content/dam/mathworks/mathworks-dot-com/compedestrianpany/events/conferences/automotive-conference-stuttgart/2018/proceedings/point-cloud-processing-using-hdl-coder.pdf
This is a task that can be done (and had been done) however by even a relatively low resolution camera even when lane lines are faded or non-existent (that was the whole selling point of the Mobileye system). Again, perception is not the problem faced by the curve problem described above that other manufacturers still face. Tesla was able to handle that on AP1 with a single camera using the EyeQ3 chip.
Joachim Langenwalter, Stellantis’ head of artificial intelligence, software, and hardware, shared the 2024 goal for the company’s Level 3 system during an online presentation on Tuesday. “The first Level 3 solution will come in 2024 before rolling out across the full portfolio in the years to come,” the executive said.
As per Stellantis’ announcement, the company is looking to deploy three platforms that are heavily geared towards vehicle software: STLA Brain, STLA SmartCockpit, and STLA AutoDrive. One of these, STLA Brain, is a Level 2 system expected to be rolled out in 2024 that is capable of being upgraded over-the-air to gain hands-free Level 3 self-driving features. Plans for Level 4 and Level 5 autonomy involve Stellantis partnering with Waymo, a company considered as the leader in autonomous vehicles today.
That looks extremely unlikely - i’ve had 2+ years of driving exactly in these kinds of conditions. So have thousands of others. If this was not one in a billion occurance we would have seen examples like this before.I'm pretty convinced Tesla's vision NN mistook the vertical glare off the headlights for two lane lines and steered in between them.
I highly doubt your take. If you look at any FSD Beta videos, you will see the path planner would happily plan a path through an oncoming lane even though the visualization shows the lane lines correctly. The lane detection is irrelevant and is not the issue here. Same thing with the examples with Blue Cruise. Given the videos shows simple cases in good weather, and we know what the Mobileye chip is capable of, it's extremely unlikely the reason why it can't stay in the lane in curves (something continually demonstrated in multiple tests, not just rare examples) is it can't detect the lanes.Mass market lidars have improved a lot from your example. But that's beside the point. Lidar can help even without identifying the lane. Consider the example below. I'm pretty convinced Tesla's vision NN mistook the vertical glare off the headlights for two lane lines and steered in between them. Lidar would have identified the large, solid object in this misidentified "lane". Lidar localizing on an HD map would have told the supervisory code that the misidentified lane was in the wrong place. This is all very valuable information, and separate from lidar itself identifying the actual lane lines.
But this whole Lidar bit is irrelevant, because the Mobileye chip Ford is using is perfectly capable of accurately telling where the lane is in the curves show in the video. If you look at the Munro video, the type of curves it has problems making with is the type you travel on AP commonly (the ones that you can easily handle doing well over the speed limit). Basically even a mild curve can be a problem.Not saying the LIDAR needs to see the lines itself. The LIDAR just tells you where the road is and therefore gives the surface to map the camera view of the lane lines onto.
I only use AP on interstates where curves can be done well over the speed limit. I guess I'll have to try it on a curvy road with hills to see how well it works. My impression from reading posts around here is that it's not all that reliable (sometimes too fast, sometimes too slow). It never seems to me that it can see more than 200ft ahead.
Of the examples you put coming, only the Luminar and Innovusion appear to have the resolution that gets closer to a low resolution camera in terms of identifying lanes. The Innoviz one you can't see the lane lines at all in the middle of the road, even up close (only can see the curbs which doesn't help you in the situations described above). The Livox might be a limitation of the visualization, but by the 30-40 meter mark (100-150 feet) you can only make them out barely intermittently.2017 called...they want their lidar back.
All cars going into production today and in the near future uses high resolution lidars, not the 4 line lidar of Scala 1. To understand the difference in resolution and spec, Luminar which will be going into several cars in 2022 has 640 lines.
Current Production Cars in 2021 with lidar:
Lucid Motors (High resolution)
Xpeng P5 (High resolution Livox lidars)
Huawei Arcfox As Hi (High resolution Huawei Lidars)
Cars releasing in 2022 with lidar (not a complete list):
Volvo (High Resolution Luminar lidar)
Nio ET7 (High resolution Innovusion lidar)
BMW IX (High resolution Innoviz lidar)
Cars releasing in 2023 with lidar (not a complete list):
GM Lyriq (High Resolution Ceptron lidar)
All the lidars are high resolution and can read lane lines, road markings, curbs, etc
Luminar Lidar output:
Innoviz Lidar output:
Innovusion Lidar output:
nio.com/cdn-static/mynio/videos/nad/nad-lidar-highway.mp4
Livox Lidar output:
From the fact that's the only one that's been massed produced at this point, and if the Mach E (which as been released a year already) was going to have it, it would likely be another Scala. And as shown above, the automakers are still using low resolution versions even from companies that have higher resolution versions available.Not sure how you are coming to this conclusion.
That looks extremely unlikely - i’ve had 2+ years of driving exactly in these kinds of conditions. So have thousands of others. If this was not one in a billion occurance we would have seen examples like this before.
That whole accident has a lot of unanswered questions - I’d not bet on anything at this point. If and when we know all the facts of the case, it will probably look pretty different from what was stated in that one post.
Tesla AP has not failed a bend/curve yet for me in 2+ years. No - I've not tried going up real mountains - but we have a lot of those curves & bends in Seattle metro (esp. on the east side in the foothills of the Cascade Range).There are plenty more examples in those videos, of the Tesla successfully navigating even sharper curves than that, but MachE fails.
Innoviz detects lane lines quite clearly and is one of the features of their perception software:Of the examples you put coming, only the Luminar and Innovusion appear to have the resolution that gets closer to a low resolution camera in terms of identifying lanes. The Innoviz one you can't see the lane lines at all in the middle of the road, even up close (only can see the curbs which doesn't help you in the situations described above).
You should dig a bit deeper. The HAP is NOT inferior to HORIZON. Its superior in every way.Edit: I dug a bit more, and the list you put is highly misleading. XPeng is using Livox HAP (their mass produced sensor), which has a 25 degree VFOV and 0.2 degree vertical angular resolution, working out to only 125 lines of resolution. No way they reliably detect lanes with that low a resolution (I'm willing to bet they will still be relying on the cameras to do that).
Empowering Xpeng P5: Livox Officially Releases HAP Lidar
HAP
The demo video you linked was of the Horizon, which claims 25.1 degree and 0.05 precision (500 lines), but the demo seems to be saying it is achieving that higher resolution by using non-repetitive scanning, which may explain why the output seems so intermittent. This is still not comparable to an actual high resolution lidar, which would be giving steady consistent output.
Horizon lidar sensor - Livox
96 lines IS high resolution lidar... Remember SDC used to use Velodyne 64 line high resolution lidar and that was all they used. Google used only one 64 line lidar for a very long time. Infact today alot still do. Most of cruise prototypes were made up of 32 lines velodyne lidars. I could keep going. To call 96 line low resolution is misinformation.The Huawei Lidar is detailed here. It's only a 96 lines (25 degree VFOV, 0.26 degree resolution), so not a high resolution lidar.
"ARCFOX Alpha S Huawei HI is the first model equipped with Huawei's three 96-channel LiDARs on the center and both sides of the front. Released in December 2020, the LiDAR has the maximum detection distance of 150m (@10% reflectivity), the field of view of 120 x 25, and the resolution of 0.25 x 0.26."
Global and China Automotive LiDAR Industry Report 2021: Application Fields, Technology and Trends, Companies and Products
Not going to continue going down the list, but clearly you would need to list the actual sensor they are using, not just the company name, as many are still obviously using low resolution lidar.
This is wrong, both livox HAP (Xpeng P5) and huawei lidars (Arcfox As and other chinese models) are being mass produced today and there are several more like innoviz that are waiting on BMW IX. But you keep changing the goal post rather than admitting you were wrong.From the fact that's the only one that's been massed produced at this point,
This isn't what you said, nor implied, you said it would "no doubt would be the type Ford would use if they opted for it".and if the Mach E (which as been released a year already) was going to have it, it would likely be another Scala. And as shown above,
no they are not. when 10-15 automakers are doing one thing, focus on the one that isn't and claim THIS IS WHAT EVERYONE IS DOING.the automakers are still using low resolution versions even from companies that have higher resolution versions available.
For the record i don't believe nor is it true that the lack of lidar is the reason for the curve problem nor the solution.I guess we'll see who is right about the Lidar contribution to the curve problem, by seeing if it will make the cars listed above magically better around curves than a Tesla. (Edit: presuming they actually select a high resolution one when the time comes, not cheap out as shown above).
As relevant to other discussion about Lidar, it uses a gen 2 Scala which has 16 lines of resolution (10 degrees VFOV, 0.6 degree vertical angular resolution):
In a world-first, Valeo’s second-generation LiDAR will equip the new Mercedes-Benz S-Class, allowing it to reach level 3 automation
Valeo SCALA 3D Laser Scanner (Gen 2)
If even Mercedes in a $100k+ S-Class is only using a gen 2 Scala when it comes out next year, this is even more evidence that if Ford opted for it, that's likely the one they would have chosen (not a more expensive higher res one).
Tesla AP has not failed a bend/curve yet for me in 2+ years. No - I've not tried going up real mountains - but we have a lot of those curves & bends in Seattle metro (esp. on the east side in the foothills of the Cascade Range).
Thats debatable. Even though they have not changed the documentation, they have released a lot of city driving specific features on AP (I'm using AP here more like EAP than current basic AP) like stop signs.(and where, of course, the system is not officially intended to work anyway since it's a 2-way road).
... It was unclear what aspect of this incident prompted the suspension...
The regulator said Pony.ai has 10 Hyundai Motor Co Kona electric vehicles registered under its driverless testing permit, and that the suspension does not impact Pony.ai’s permit for testing with a safety driver.
Yes, though I think they often have a "safety passenger". The accident was odd, it ran onto a median and hit a sign on a curved section of road. It's almost as if it tried to go in a straight line from one waypoint to another and cut across the curve. Or maybe it tried to lane change into a non-existent lane. It doesn't seem any other cars were involved, just a basic perception/localization error. Not something you'd expect from a system approved for driverless testing.It sounds like when the accident, there was no human safety backup driver present.
...basic perception/localization error. Not something you'd expect from a system approved for driverless testing...
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a recklessly behaving road user |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a recklessly behaving road user |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a recklessly behaving road user |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a recklessly behaving road user |
Disengage for a recklessly behaving road user |
Disengage for a hardware discrepancy for which our vehicle's diagnostics received a message indicating a potential performance issue with a hardware component of the self-driving system or a component of the base vehicle |
Disengage for a hardware discrepancy for which our vehicle's diagnostics received a message indicating a potential performance issue with a hardware component of the self-driving system or a component of the base vehicle |
Disengage for incorrect behavior prediction of other traffic participants |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a hardware discrepancy for which our vehicle's diagnostics received a message indicating a potential performance issue with a hardware component of the self-driving system or a component of the base vehicle |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a recklessly behaving road user |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a recklessly behaving road user |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for incorrect behavior prediction of other traffic participants |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for incorrect behavior prediction of other traffic participants |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for incorrect behavior prediction of other traffic participants |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for incorrect behavior prediction of other traffic participants |
Disengage for incorrect behavior prediction of other traffic participants |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a recklessly behaving road user |
Disengage for a recklessly behaving road user |
Disengage for a hardware discrepancy for which our vehicle's diagnostics received a message indicating a potential performance issue with a hardware component of the self-driving system or a component of the base vehicle |
Disengage for a hardware discrepancy for which our vehicle's diagnostics received a message indicating a potential performance issue with a hardware component of the self-driving system or a component of the base vehicle |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a hardware discrepancy for which our vehicle's diagnostics received a message indicating a potential performance issue with a hardware component of the self-driving system or a component of the base vehicle |
Disengage for a hardware discrepancy for which our vehicle's diagnostics received a message indicating a potential performance issue with a hardware component of the self-driving system or a component of the base vehicle |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for a recklessly behaving road user |
Disengage for a hardware discrepancy for which our vehicle's diagnostics received a message indicating a potential performance issue with a hardware component of the self-driving system or a component of the base vehicle |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a recklessly behaving road user |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for incorrect behavior prediction of other traffic participants |
Disengage for a perception discrepancy for which a component of the vehicle?s perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for adverse weather conditions experienced during testing |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a recklessly behaving road user |
Disengage for adverse weather conditions experienced during testing |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for incorrect behavior prediction of other traffic participants |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a perception discrepancy for which a component of the vehicle's perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for adverse weather conditions experienced during testing |
Disengage for adverse weather conditions experienced during testing |
Disengage for a perception discrepancy for which a component of the vehicle's perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle's perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle's perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle's perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle's perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for a perception discrepancy for which a component of the vehicle's perception system failed to detect an object correctly |
Disengage for a perception discrepancy for which a component of the vehicle's perception system failed to detect an object correctly |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
Disengage for unwanted maneuver of the vehicle that was undesirable under the circumstances |
What ?!?Pony.ai backed by Toyota gets its driverless license in California suspended for hitting a divider.
What ?!?
I was promised that with Lidar this would NEVER, EVER happen !!
That's what people are saying above in response to the whole discussion about curves, that lidar does not solve the problem.Lidar doesn't remove the possibility of this happening.
What it does do is reduce the possibility that it's a sensor detection issue.
It doesn't eliminate the possibility that logic downstream won't have mistakes.