12.6 - Startover
6th edition of the 12th year of SmartDrivingCars eLetter
A Waymo driverless car didn't see the person on a bike it hit, company says
J. Serrano, Feb. 8 "One of Waymo's driverless cars collided with a cyclist in San Francisco this week because it apparently didn't see the human until it was too late.
The incident occurred on Tuesday at a four-way intersection while the self-driving vehicle was at a complete stop, Waymo, which is owned by Google, told local outlet ABC7. The car was waiting for a large truck to pass and drove into the intersection when it was its turn. That was when it hit the cyclist, who was riding close to the tail end of the truck. The unnamed victim suffered no serious injuries, according to Waymo, and walked away on their own."...
Hmmmm… OK, a 4-way stop. Traffic laws dictate that all vehicles must stop, and whoever stopped first then proceeds first after the others have stopped. So, if the Waymo was stopped to let the truck go, the other human operated object behind the truck (the bicycle) got to its stop sign after the Waymo, and hence should have allowed the Waymo to proceed. If that's the case, then, again: human driver misbehavior contributed to this mishap.
Waymo, however, is not off the hook entirely. Apparently, Waymo was not able to sufficiently detect what was following behind a truck and it struck that entity. That is not good irrespective of that entity being identified as a human-operated bicyclist, scooter or even some small object that the truck was pulling. Since we should all be cooperating on safety, irrespective of what anti-trust law or government regulations might or might not permit, it is imperative that Waymo makes all the data and simulation results of this incident available to everyone. And it seems as though they are doing just that - it looks like one important lesson of the Oct2 GM/Cruise SF incident is yielding value. Making this information available will help everyone avoid not only cyclists surfing behind trucks but also avoid objects that a truck might be towing.
That accomplished, we'll now have to wait to see what the press/public response is to this incident. If this characterizes Waymo as being insufficiently safe to operate in San Francisco or any other jurisdiction, then Waymo and much of the Driverless car industry needs to Start Over! They are unable to solve the problem they have stated as the problem they are singularly focused on solving: safety. They need to read "The Book" (it's only $100, while they've spent $300 Billion not understanding what they are trying to fix.)
More.. If one wants to look at human behavior at stop-controlled intersections one may start with M. Woldeamanuel, fall 2012: "Stopping Behavior of Drivers at Stop-Controlled Intersections: Compositional and Contextual Analysis".
Also see Andrew Hawkin's take on this incident: Waymo driverless car strikes bicyclist in San Francisco, causes minor injuries. Alain
Waymo driverless car vandalized, set on fire in San Francisco's Chinatown
Staff, Feb. 11, "A Waymo autonomous vehicle was set on fire Saturday night after it was vandalized by a group of people in Chinatown, San Francisco firefighters said.
The vehicle was traveling on Jackson Street, between Stockton and Grant, about 9:25 p.m. when it was surrounded by about 10 to 15 people, Lt. Mariano Elias with San Francisco fire said.
Several social media videos showed a group of people vandalizing the self-driving car. While another video showed the Waymo vehicle going up in flames…."
Hmmmm… I had titled this edition of the SDC eLetter "Start Over", before this happened. I'm speechless, but watch the ZoomCast and Read the Book!!!. Alain
Just Published!!! Go to Amazon.com… You can still be first on your block to have one J.
https://www.amazon.com/Real-Case-Driverless-Mobility-Vehicles/dp/0443236852/ref=sr_1_1
SmartDrivingCars
ZoomCast 357 / PodCast 357 Waymo vehicle burned in S.F.
F. Fishkin, Feb. 11, "The burning of a Waymo vehicle in S.F. may be shocking, but not surprising according to Princeton's Alain Kornhauser. In the latest Smart Driving Cars with co-host Fred Fishkin, Alain again points to The Real Case for Driverless Mobility (new book co-authored with Michael Sena). Plus..Waymo collision with a cyclist investigated, Apple Vision Pro headsets behind the wheel (Tesla), MOVE mobility for rural areas and Ken Pyle's Texas EV rental adventure.
- 0:00 open
- 0:45 Waymo vehicle burned in S.F.
- 22:40 Waymo vehicle cyclist collision under investigation
- 29:00 Apple Vision Pro headset being worn behind the wheel by some Tesla drivers?
- 40:20 Is MOVES style mobility for rural areas disruptive?
- 52:20 Ken Pyle's Texas EV rental adventure
- 59:50 The Real Case for Driverless Mobility book in perspective"
Stop Wearing Vision Pro Goggles While Driving Your Tesla, U.S. Say
J. Jimenez, Feb. 6, "Videos being shared across social media this week depict an almost dystopian, futuristic scene: drivers of Teslas in Autopilot mode while wearing Apple Vision Pro headsets, seemingly unaware of the road in front of them.
The videos led federal transportation officials to issue warnings.
But are people really mindlessly riding around in Teslas in Autopilot mode, wearing Apple's futuristic new goggles? Or is it all just a bit? Part of a never-ending cycle of people doing silly things for clicks, likes, views and clout?
The new goggles have a feature that merges digital apps and one's surroundings into one immersive space, and videos of people wearing them in strange settings have started to crop up across the internet since they were released on Feb. 2.
Several of the videos taken in cars appear staged, and in many, it is clear that someone other than the driver is recording. The videos are not widespread. Still, they seemed reckless enough for Pete Buttigieg, the transportation secretary, to weigh in on social media.…."
Hmmmm… Wow, the NY Times and The US Department of Transportation have realized that some drivers misbehave when driving on public roads. I've written on these pages that "greater than 90% of what the NY Times and US DoT often call car "accidents" in fact involve human "mis-behavior" and should be called out as such and not as "mere accidents".
Moreover, NHTSA and Tesla should "cut a deal" by which NHTSA would require Tesla to disable its Advanced Driver Assistance Systems (autoPilot, FSD and all other automated driver assistance systems that it has on its cars) whenever the driver monitoring system detects that the driver is misbehaving in the explicit, even fine print, use of any of these systems. First offenders could reinstate their use by viewing some on-line use-training video, second offenders might need a note from their insurance company or their bank. A third strike might involve a substantial fine or permanent ban. Such cascading penalties are not a new concept. What is new here is that NHTSA would absorb the customer kickback and Tesla could simply wash their hands and claim… "They made me do it. Talk to them about reinstatement".
Driver misbehavior is not limited to Tesla drivers. All OEMs have a few of their customers mis-using the product that they've sold them. One could make the above practice mandatory for all, for example not allowing the car to travel at speeds of more than "9-over" (or whatever the number). Everyone would be safer. Over-the-Air (OtA) updating of software makes such cooperation trivial for Tesla, rather than exceedingly expensive for all the other OEM that haven't embraced OtA. What an opportunity for Tesla to become substantially safer and more customer conscious than every other OEM. Think about it. Alain
Is MOVES-style Mobility for Rural Areas Disruptive? The Tuscarora Reservation Opportunity…
A. Kornhauser, Slides used in his class Feb. 7, "…Total Addressable Market is 3,165 personTrips per day…,".
Hmmmm… (and be sure to watch the video on the last slide). Not that much, but the area is really rural, and many of those rides are really needed. What these slides and the end video reveal explicitly is that 1,000 of these rides could actually be given most days at a very good demand-responsive level-of-service (less than 5 minute wait, direct to destination with minimal circuity in the few ride-share opportunities) achieving vehicle productivities of >40 personTrips/Vehcicle-day (>12,500personTrips/Vehicle year, implying that capital costs for such a system could be financed by less than $2/personTrip and break-even fares of >$3/ride. Now that's what I call affordable high-quality mobility. All we need is a Vehicle that works and we or almost any franchise could offer Mobility within even some of the most rural settings in the nation. Now that's interesting!! Alain [and thanks, Bryce, for getting the analysis system and simulation to work!!]
EV Charging Batters My Self Esteem
K. Pyle, Feb 10, "It is two decades later and the technology overlords continue to batter my self-esteem. In 2005, it was the lowly PC that beat me into submission. In 2024, it is EV charging and the complications of trying to juice up in Texas that almost had this author in tears…."
Hmmmm…. Enjoy!! I'm still struggling with my iPhone and please turn off all that background ChatGPT stuff. The latency that it has introduced is taking us back to "1984".
Alain