OxBrew
Active Member
40,000 traffic fatalities a year in the US on the current system. So, maybe we do need to talk to each other a bit more?
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Autonomous Car Progress
- Thread starter stopcrazypp
- Start date
-
- Tags
- Autonomous Vehicles
Robotaxis go on strike for better pay and working conditions.
diplomat33
Average guy who loves autonomous vehicles
Presumably, the city or the tech company would maintain it. But IMO, Vehicle to grid is not going to happen. AVs are not going to be built to rely on a grid in order to drive autonomously. That is silly. AVs are built with sensors and software to make sense of the world and make their own driving decisions. The AV software will be good enough that AVs won't need to rely on a grid to drive.
You didn't read (or quote) the entire sentence. What I said was "Car-to-infrastructure may be what's needed to move consumer L3 autonomous systems built on today's platforms... to an L4 system, with the driver having ample time to takeover when the infrastructure communications indicate something ahead is amiss." In other words, while Waymo and others with ample sensor suites can support L4, the minimal cameras, radar, and sonar on consumer systems like, e.g., my Tesla M3, will probably only ever support L3, with the driver having to stay engaged and be ready to take over immediately when some exception arises. V2V and V2I communication systems may, however, enable these L3 owner systems to become L4, since the car could potentially have more foreknowledge of a change in conditions and institute a fallback in a reasonable amount of time, e.g., give time for the driver to grok in and take control or slow the car and pull onto the shoulder.
Daniel in SD
(supervised)
I see, I didn't really consider that V2V. What types of "exceptions" are you envisioning here?
Or, what types of collisions would such a system prevent?
Here's one example: imagine that a current autonomous system relies heavily on detailed mapping data (not like HD Lidar) for lane selection and navigation through intersections in metropolitan areas. Imagine that without up-to-date mapping data the performance of the system in lane recognition is very capable but not to a confidence level where the manuf. thinks it qualifies for L4. So maybe with mapping data the system is L4, and without mapping data it's L3.
Now imagine that road crews come out to restripe the pavement or add a lane. They could utilize a V2I beacon to tell approaching autonomous cars that their mapping data may be wrong and to take appropriate remedial action. The L4 system may, if a driver is present, alert the driver of the approaching anomaly and require the driver to pay attention as it traverses the construction zone. Or, if no driver is present, it could find an alternate route or pull over to the side of the road until administrative personnel could provide remote instructions for what remedial actions to take.
More media coverage on Cruise mishaps:
www.kron4.com
Cruise driverless vehicles involved in 3 separate SF traffic incidents Thursday
SAN FRANCISCO (KRON) — In separate incidents on Thursday night, three Cruise driverless vehicles created disruptions and blocked traffic on the streets of San Francisco, with one vehicle appe…
www.kron4.com
diplomat33
Average guy who loves autonomous vehicles
If the AV needs up-to-date maps just to do something basic like lane keeping at L4 reliability, then it is not good L4. You can't have a system that downgrades to L3 every time it encounters a change in the map. That is not going to work. AVs need to be able to do lane keeping reliably without maps precisely so that it can handle cases when the maps are wrong. The manufacturer needs to make their lane keeping more reliable without maps. And I believe in HD maps. But I am a proponent of HD maps as a prior, as an aid to enhance higher safety, not as a requirement to do basic lane keeping.
You could have V2I beacons for extra redundancy but it cannot be the primary system IMO. You don't want AVs that are dependent on V2I beacons to be able to handle changes in the map. That is not going to work. What if the beacons don't work or the construction zone does not use the right beacons? AVs need to be able handle changes in maps on their own. For example, I know Waymo and Mobileye are able to detect changes in the map on their own using their cameras and can automatically update their maps and share those map changes with the fleet. That is the best way to handle things IMO.
FloridaJohn
Member
V2X is dead. Frequency spectrum was allocated to this function in 1999 and not a single vehicle manufacturer has done anything with it. The spectrum is being reassigned.
arstechnica.com
Court rules FCC is allowed to reassign 5.9 GHz bandwidth, killing V2X
Bandwidth was set aside in 1999, but V2X has been an abject failure.
arstechnica.com
We already do somewhat have V2V communication. It comes in the form of traffic data. Routing decisions are being made based on this information. While this data is nowhere near as fine grained as what I believe you're talking about, the concept is the same.
The problem with 6 GHz was that it does not propagate far enough to be effective (think 5 GHz WiFi). It might work for vehicles gathered at the same 4-way stop, but not far enough to really affect traffic flows.
Daniel in SD
(supervised)
The thing is I've never seen a reported AV collision caused by road changes so I'm skeptical that this would be useful.
Of course Tesla does not use HD maps so this system would be useless to them as well.
Without sparking (yet another) debate of the differences of L2, L3, and L4, when it comes to autonomous systems driven primarily by machine learning, I don't think there are necessarily bright lines. In machine learning, everything comes down to confidence levels. A neural network won't tell you that something is, e.g., a dog or a cat. It will tell you that it thinks there is an 92% chance it's a dog, and a 5% chance it's a cat.
A system may be designed and have all functionality to be L3, but if aggregate confidence levels aren't generally high enough, it may behave erratically and be deployed by the manufacturer as an L2 driver assist, as is the case of with Tesla FSDb. Similarly, a system may be designed as an L4 system, but instead of geographic, its limited domain may be environmental - e.g., visibility, reliability of map data, lighting, etc. The system may only operate at the confidence levels to be L4 when the environment is ideal. It that a "good L4?" Maybe not, but I think you would agree that the thing that makes Waymo L4 over Tesla's L2 is not necessarily design goals of the system at this point but availability of data (sensors, maps, GPS, etc.). Tesla having limited sensors and low resolution mapping has hampered it's ability to become a true L3 system. And, IMO, without additional sensors and/or data, it can never be L4. Accordingly, that's why I say V2V and V2I could provide additional data to these types of systems so that their confidence levels, when such data is available, are high enough for these system to operate at L3 (or maybe L4).
Will it happen? I doubt it. But could it add to the confidence levels of the NNs and thus the performance of the system? Absolutely.
The problem with something like Waze is that there is tremendous latency, so while it may help for routing around problems on a long trip, it can't really add much to autonomous vehicles. But imagine that you are driving down the highway and it starts to rain hard and visibility is reduced. The system may be designed to fallback to L2 in those conditions. However, if the car was in direct communication via V2V with one or more cars in front of it on the same road travelling in similar vectors, that may be enough information for it to have the confidence levels to continue to operate at L3 despite the reduction in visibility.
You think that car manufacturers will invest in a complex V2V system just to allow cars to drive more closely together and faster in fog or heavy rain? I think that's not worth the trouble.
The first car in a convoy does not gain anything and has to drive slowly anyway. So there is no advantage at all.
And as long as the other cars are visible, I fail to see any need for V2V communication.
The only exception I can foresee is car convoys where the lead car takes most of the responsibility, and the following drivers can rely on the leader and, for example, sleep. Could be interesting for trucks.
The first car in a convoy does not gain anything and has to drive slowly anyway. So there is no advantage at all.
And as long as the other cars are visible, I fail to see any need for V2V communication.
The only exception I can foresee is car convoys where the lead car takes most of the responsibility, and the following drivers can rely on the leader and, for example, sleep. Could be interesting for trucks.
Who said anything about speed or closeness? It’s about whether the car has enough information to have the confidence to drive autonomously or needs to fallback to driver monitored or driver control.
diplomat33
Average guy who loves autonomous vehicles
Big news for Mobileye SuperVision:
SuperVision coming to 3 additional brands under Geely and 2 new Zeekr EV models next year:
Mobileye also expected to roll out full SuperVision capabilities by the end of this year to Zeekr001:
SuperVision description:
www.intel.com
SuperVision coming to 3 additional brands under Geely and 2 new Zeekr EV models next year:
Mobileye also expected to roll out full SuperVision capabilities by the end of this year to Zeekr001:
SuperVision description:
Geely Holding Group Expands Mobileye Collaboration
Three additional Geely brands to leverage Mobileye SuperVision for advanced ADAS, building off the successful launch of the ZEEKR 001.
www.intel.com
diplomat33
Average guy who loves autonomous vehicles
Yes, I agree with all that. My earlier point was just that AVs should not be built to depend on V2I for L4. But I am ok with V2I as an aid to help systems with fewer sensors achieve better reliability. I do think Waymo's sensors+maps are good enough for reliable L4. I don't think Waymo needs V2I. But systems like Tesla could benefit from V2I.
diplomat33
Average guy who loves autonomous vehicles
Similar threads
- Article
