P. Loeb, May 16, "…Sponsor Cherelle Parker says the cameras will photograph any car going more than 11 miles per hour over the speed limit…" Read more Hmmmm… I really don’t understand. What is the meaning of the word limit ? (Hint…. "the utmost extent")
So for humans a "speed" limit is actually a "Speed +10" limit. That mean I can set my Cruise Control to "Speed Limit" +10 and I’ll be just fine. Does that also mean that I can code my driverless car "to do +10"??? If not, then why does a person capable of getting a driver’s license get to go faster than a person who can’t get a driver’s license who is relegated to be driven by an autonomousTaxi (aTaxi) that is mandated to drive at a slower speed???? (Please don’t tell me it is because the accuracy of the speed sensor is not precise (aka reliable enough). May I use that excuse in my aTaxi code?) This is a serious question! There needs to be a level regulatory (rules of the road/traffic laws) playing field established for aTaxis and human drivers. This is NOT easy (but it could be as simple as:
SpeedLimit(aTaxi) = SpeedLimit (Humans) + 10
StopSign(aTaxi) = SropSign(Humans) +RollOnThrough if no one is around
RedLight(aTaxi) = Redlight(Humans) + 3 more cars after the yellow, except in Boston where 5 more car after the yellow… Alain
F. Fishkin, May 25, " Say what you mean and mean what you say. When it comes to traffic laws, that’s the way autonomous vehicles see it. What about the rest of us?… asks Princeton’s Alain Kornhauser. Tune in as he joins co-host Fred Fishkin on that issue, the future of Tesla, the USPS tests self driving trucks and more on GM’s Cruise and Google’s Waymo." Just say "Alexa, play the Smart Driving Cars podcast!" . Ditto with Siri, and GooglePlay … Alain
C. Gunther, May 24, "The National Highway Traffic Safety Administration is setting the stage for addressing the safety of self-driving cars: how to regulate them and what regulatory changes may be to adequately assess their safety.
In a Federal Register notice to be published May 28, the NHSTA points out as self-driving cars could reduce the number of vehicular crashes and injuries, it plans to develop a new "comprehensive strategy" to update safety standards for automated vehicles "while addressing regulatory barriers to the compliance verification of these vehicles."… " Read more Hmmmm… Read the notice carefully!!! This is important. Unfortunately, NHTSA continues to use the SAE levels and doesn’t properly address Driverless, which should be created as a new mode and taken out of NHTSA’s domain with the creation of a new Federal Driverless Motor Vehicle Safety Administration (FDMVSA) focused on establishing FDMV Safety Standards (FDMVSS) focused on collision avoidance as well as collision mitigation, not visa versa as is NHTSA’s heritage. This is important!!! NHTSA has enough to do in dealing with vehicles with drivers in the loop because the driver is a major part of the problem. Another entity should regulate vehicles where that part of the problem is irrelevant. Alain
M. Cole, May 23, "Two agencies within the U.S. Department of Transportation are seeking public input on removing regulatory barriers to allow the integration of autonomous vehicles onto public roads. The Federal Motor Carrier Safety Administration and the National Highway Traffic Safety Administration announced Wednesday they will publish advance notices of proposed rulemaking “to ensure that all potential approaches are fully considered” while moving forward with autonomous vehicle regulations.
FMCSA’s notice will seek public comment on questions regarding several regulatory areas to better understand how changes to its rules can account for differences between human drivers and automated driving systems (ADS). The agency says the questions focus on requirements of human drivers; CDL endorsements; hours of service rules; medical qualifications; distracted driving; safe driving; inspection, repair and maintenance; roadside inspections; and cybersecurity.
“FMCSA is hoping to receive feedback from commercial motor vehicle stakeholders and the motoring public on how the agency should adapt its regulations for the development of increased automated driving systems in large trucks and buses,” said FMCSA Administrator Ray Martinez. “We know that while many of these technologies are still in development, it is critical that we carefully examine how to make federal rules keep up with this advancing technology.”…
NHTSA will seek comment on identifying and addressing regulatory barriers to the deployment of ADS vehicles posed by certain existing Federal Motor Vehicle Safety Standards (FMVSS). …See above article… =The agency also wants to hear from the public on different approaches that could be used to measure compliance with the FMVSS for vehicles without conventional controls, including steering wheels and brake pedals…." Read more Hmmmm… Read the notices carefully!!! FMCSA’s and NHTSA’s This is important.! Alain
R. Mitchell, May 22, "Morgan Stanley stock analyst Adam Jonas told clients Wednesday that if Tesla’s troubles continue, it’s unlikely to be acquired by a tech company, an auto company, or someone from China. But it could end up merging with SpaceX, another Elon Musk company.
A recording of the 55-minute conference call with clients was leaked online. The meeting was held two days after Jonas, a longtime Tesla bull, put a “worst-case” price of $10 on Tesla stock, although he didn’t change his current $230 target price.
While the SpaceX option is “an admittedly fantasy case,” Jonas said it’s more likely than any established player buying the debt-ridden electric car maker.
Tesla is moving “from a growth story to a distressed credit and restructuring story,” Jonas said…. But if demand doesn’t pick up substantially and the capital market for Tesla dries up, the company will need to pursue other options, Jonas said.
That includes a possible acquisition, but Jonas said it’s unlikely. Tech companies such as Apple are unlikely to be interested, he said, to avoid the financial risk of “being involved in owning a business where occasionally a car catches on fire, takes down a building, accidentally kills a pedestrian or a passenger … the auto industry has an ugly side to it, the roads are very dangerous.” Tech companies working on driverless cars “realize that the autonomous race is more of a marathon” that might take 10 or 20 years for full deployment. …
Auto companies probably would not want to take on Tesla’s substantial debt and infrastructure, Jonas said. With 49,000 employees, Tesla takes in $500,000 revenue per worker, Jonas said, while Ford and General Motors pull in $850,000 per worker. “They built this hulking infrastructure to support more like 1 million cars a year, not 350,000 cars year,” Jonas said. And most analysts don’t think Tesla will build that many cars in 2019.
The company’s $13 billion in gross debt is the biggest impediment, Jonas told his clients. “If someone could get access to Tesla’s assets and have it with the right number of employees and no debt, there is asset value there,” he said. … Chapter 11?… …
A possible sale to Chinese investors is out, he said. Given the current trade war, passing U.S. government review at this time is would be impossible.
That leaves SpaceX, the private rocket company that Musk also founded. The latest public information available, from 2016, shows “Mr. Musk’s trust” owns 54% of the company, and has voting control of 78% of the stock…. " Read more Hmmmm… Wow!!! A space company bailing out a fundamentally terrestrial company. We have come a long way since Oct. 4, 1957, 7:28pm GMT. One view of how much we have spent is at https://stuffin.space/. And these are only the stuff still in orbit. A lot has decayed and de-orbited, including Sputnik 1. Who would have thought? Mow SpaceX puts 60 in orbit with one rocket. Alain
S. Szymkowski, May 24, "Cruise Automation is humming along as General Motors works to beat the industry to a proper commercial platform for self-driving cars. The latest bit of Cruise propaganda to show off how smart its cars are becoming focuses on unprotected left-hand turns. The maneuver is characterized by any time a driver needs to make a left as oncoming traffic flows and pedestrians cross in the path of turning vehicles. The kind of turn is even more difficult in metropolitan areas, like San Francisco where Cruise continues to test its technology….
“In an unpredictable driving environment like San Francisco, no two unprotected left-turns are alike,” Kyle Vogt, the company’s president and chief technology officer, said in a release. “By safely executing 1,400 [left turns] regularly, we generate enough data for our engineers to analyze and incorporate learnings into code they develop for other difficult maneuvers.”… " Read more Hmmmm… See video!! We can call it propaganda or we can simply be impressed by the fact that it may just work. 1,400 left turns is impressive; however, every day there are about 800,000 car trips that execute say and average of 2, 3 (?) unprotected left turns per trip (pick a number). That’s > 1 million per day. Most, say 99% are trivial, but that still leaves 10K that are challenging, 99% of which are no-problem. That still leaves 100 per day that may well be WTx? And that’s on a daily basis. However, we are getting there and if we go through those turns without trying to power slide through them, we actually may be able to do this. 🙂 Alain
K. Lang, May 23, "General Motors Co. is running into opposition to its petition to federal regulators for permission to put up to 5,000 driverless cars — without steering wheels or control pedals — on public roads.
In comments submitted to National Highway Traffic Safety Administration, groups that represent car dealers, insurers and road safety advocates took issue with the Detroit carmaker’s request to put 2,500 "driverless zero‐emission autonomous" vehicles on the road annually for a two-year period beginning this year….
Thomas Karol, general counsel for federal issues for the National Association of Mutual Insurance Companies, which lobbies for property and casualty insurance providers, said NHTSA "… Read more Hmmmm… Shame on Insurance! Totally self-serving. If you were a true safety advocate, you’d be lobbying on the other side of the fence. What you fear is that the carnage that you currently insure will disappear and you’ll be out of business. So UGLY. Alain
C/ Linder, May 24, "Self-driving car startup Aurora Innovation is revving up against the competition. The Lawrenceville-based company announced this week it has acquired Blackmore, a Bozeman, Mont.-based firm that makes lidar — a sensor that helps the vehicles “see” — to shore up development of its own self-driving system.The deal makes Aurora the latest in the industry to bulk up its in-house computer vision tech. Other industry players, which in Pittsburgh include Uber, Aptiv and Argo AI, have been gobbling up lidar companies like wildfire….
Details of the financial transaction were not disclosed. As part of the deal, Aurora has opened a fourth engineering center in Bozeman, in addition to keeping offices in Pittsburgh, San Francisco and Palo Alto, Calif. Having in-house lidar, which is the most important sensor for self-driving computer vision, is far more cost-effective than relying on off-the-shelf units….
In October 2017, Argo AI in the Strip District acquired Princeton Lightwave, a New Jersey-based lidar company. The company’s technology will aid Argo’s virtual driving system, increasing the accuracy of object detection in unusual and challenging circumstances caused by poor weather. …weather???… …" Read more Hmmmm… See video. Interesting and in line with the discussions during Workshop 7 of the 3rd Princeton SDC Summit. Congratulations Steve Crouch! Alain
J. Stoll, May 24, "Here’s what we know about autonomous vehicles: They’ll be safer and smarter than the cars parked in our driveways today. Anyone who speculates beyond that is just, well, speculating. It wasn’t long ago that a top automotive supplier was talking about providing fully driverless systems for production vehicles by 2019. Uber thought it would have 75,000 autonomous vehicles operating by this year. Elon Musk pointed in to a similar horizon in 2014 when a group of Journal editors and reporters asked him when he thought an entirely autonomous Tesla would hit the road.
If building a robocar were simply about figuring out how to best arrange a box of sophisticated parts, these predictions may have come true. Most of the necessary components—sensors, cameras, chips, those bulky lidar units that sit on top of the car—have been around for awhile. Any car maker or parts supplier worth its salt could figure out how to gin up a remote-control SUV.
But it takes gobs of engineers, data, software, patience and cash to teach that 4,000-pound vehicle to think for itself…" Read more Hmmmm… Yup!!! But the value to society that can be delivered has NOT dissipated. It is just going to take blood, sweat, tears and money, not lipstick to accomplish it. Silicon Valley and Wall Street are going to need to earn this one. Alain
K. Barry, May 22, "Tesla added the lane-changing update to its Navigate on Autopilot feature last month as part of a promised upgrade to the package of driver assist features. We first reviewed Navigate on Autopilot in November and found it technologically impressive. But we also raised concerns about its performance in heavy traffic.
To enable the new feature, a driver must first change the system’s settings, essentially giving the car permission to change lanes on its own. The driver can cancel an automated lane change that’s in progress at any time by using the turn-signal stalk, braking, or holding the steering wheel in place. .Excellent…
In practice, we found that the new Navigate on Autopilot lane-changing feature lagged far behind a human driver’s skills. …Not at all surprising for most "driver’s skills", for which there is a wide distribution of capabilities… The feature cut off cars without leaving enough space, and even passed other cars in ways that violate state laws … As is done regularly by many/most human drivers… , according to several law enforcement representatives CR interviewed for this report. As a result, the driver often had to prevent the system from making poor decisions.
“The system’s role should be to help the driver, but the way this technology is deployed, it’s the other way around,” says Jake Fisher, Consumer Reports’ senior director of auto testing. “It’s incredibly nearsighted. It doesn’t appear to react to brake lights or turn signals, it can’t anticipate what other drivers will do, and as a result, you constantly have to be one step ahead of it.”…All true. The system needs to be better … "“In essence, the system does the easy stuff, but the human needs to intervene when things get more complicated,” Fisher says. …At least it does something .
…“Tesla is showing what not to do on the path toward self-driving cars: release increasingly automated driving systems that aren’t vetted properly,” he says. “Before selling these systems, automakers should be required to give the public validated evidence of that system’s safety—backed by rigorous simulations, track testing, and the use of safety drivers in real-world conditions.” … …I hate to be supporting Tesla, but there is no evidence that Tesla hasn’t done what is suggested here and is now at the next stage…
Autopilot has been engaged during at least three fatal crashes in the U.S., according to the National Transportation Safety Board. The latest one occurred in March, when the driver of a Tesla Model 3 was killed after the vehicle struck the side of a semitrailer in Florida. The driver turned Autopilot on 10 seconds before the collision, according to the NTSB’s preliminary findings released last week. "…" Read more Hmmmm… With all of its wisdom, CR fails to call out the Automated Emergency Braking (AEB) system as the system of failure in Teslas and essentially all other cars on the road. When is CR going to call out the real nemesis here? Alain
May 2019, "Mission accomplished. When UK Autodrive got underway in October 2015, it was the largest trial of connected and self-driving vehicles ever to have taken place in the UK. Back then, the concept of autonomous cars would still have struck many people as the stuff of science fiction. And yet, in the space of just three years, we have gone from concept to reality – delivering on our promise to trial fully connected and self-driving vehicles on UK streets…." Read more Hmmmm… For what it is, this is a good report; however, it seems more like GW’s Mission Accomplished. Yes, progress has been made, but if the objective was "connected", that hasn’t progressed, and if it was "self-driving", then what has been accomplished isn’t much more than what Tesla and GM/CT6 sell today. Driverless in everyday conventional situations is as GW’s Mission Accomplished. Yes there has been enormous progress, but we are still at the "Gentlemen, start your engines", if we’re even at the start line of actually providing mobility to anyone, let alone the people whose quality of life could be most enhanced. (to which there is no mention in this report (or if it is, I missed it. Please let me know so that I can correct this post.) Alain
J. Bhuiyan, May 19, "Former Uber chief executive Travis Kalanick once blamed the high cost of each ride on the "other dude in the car" – the driver.
It has long been on the ride-hailing company’s road map to someday replace its vast workforce of contractors with self-driving vehicles, lowering fares and increasing profits. But even Uber acknowledges that that won’t happen any time soon….
"In the aftermath (of the crash), Uber really had all the signs that they were cutting corners and were trying to push it to market faster than it should be going," said Paul Sagawa, an analyst at SSR. "The investigation around that fatality in Phoenix reveals a company whose technology is well behind, years behind Google’s in terms of its ability."
Efforts to automate all or part of these transactions may be ongoing, but Uber and other companies have come around to the idea that they’ll likely continue to rely on drivers for much of it. Even with self-driving, Uber says that "there will be a long period of hybrid autonomy, in which autonomous vehicles will be deployed gradually against specific use cases while Drivers continue to serve most consumer demand". Read more Hmmmm… Reality can’t be worth $xxB. Alain
I. Boudway, May 21, "The United States Postal Service is going to put mail on self-driving trucks. Starting on Tuesday morning, letters and packages moving between Phoenix and Dallas will travel on customized Peterbilt trucks run by TuSimple, an autonomous startup based in San Diego. There will be five round trips between the two cites, with the first haul leaving from Phoenix this morning. It’s the first time that the Postal Service has contracted with an autonomous provider for long-haul service…
For now, however, TuSimple will have a safety driver behind the wheel for the 1,000-mile trip between Phoenix and Dallas, as well as an engineer in the passenger seat monitoring the autonomous systems. In the future, the startup aims to provide “depot-to-depot” service without drivers. (Alain added Bold) …" Read more Hmmmm… Another conception, but we need to get to an actual live birth. Pay attention to the bold print before you get too excited. At least this is termed "Self-driving" and not "Driverless". It is actually "worse" than self-driving in that there is " as well as an engineer in the passenger seat " . I hope that the "engineer" can also serve as a tandem driver to get around hours of service rules. Alain
F. Lambert, May 22, "Right after helping Tesla (TSLA) raise over $2 billion, Wall Street is turning on the automaker with a series of negative analyst notes predicting a lot of downside for Tesla investors.
Yesterday, Morgan Stanley analyst Adam Jonas, who was once called Tesla’s biggest cheerleader, was out with a new note for clients. In the note, the analyst actually kept his main price target for the stock at $230 and even noted that he sees a bull-case valuing Tesla at $391, but the media focused on his “bear case”, which included a $10 a share valuation.
He wrote: “Demand is at the heart of the problem. Tesla has grown too big relative to near-term demand, putting great strain on the fundamentals.” The note put pressure on Tesla’s stock yesterday, but it erased most of its losses by the end of the day."…
It always surprises me that these guys have such a big impact on markets and stocks like Tesla’s considering they have been consistently wrong.
I mean Jonas predicted that Tesla would be so late that the Model 3 that it would only arrive in “2018” and volume deliveries wouldn’t start until 2019. Tesla ended up delivering almost 150,000 Model 3 units in 2018. It’s just one of many examples of those supposedly top Wall Street analysts getting things wrong when it comes to Tesla.
I wouldn’t base any investment decision on what they are saying, but it looks like many are since they are moving the market. " Read more Hmmmm… But a 150,000 to market in 2018 does not qualify for a market cap greater than Ford’s. Alain
F. Lambert, May 23, "Tesla dominates other premium automakers in California, pushes EV market share over 5%…According to the same registration data, Tesla delivered 15,805 Model 3 vehicles in the first quarter 2019-more than the next 4 competitors combined:" Read more Hmmmm… That is really impressive, but again, is the stock price simply too high. Alain
F. Lambert, May 23, "… Earlier this month, a Tesla owner claimed that Autopilot stopped for a rabbit on the road caught on a dashcam video. At the time, we noted that the accuracy of the owner’s claim is unclear and that a visualization of what the Autopilot can see, like hacker verygreen has been producing, would be useful in this situation.
While green couldn’t produce a rabbit out of his hat, a goose did appear on the road while the Autopilot data was being collected and he produced the following visualization using that data:…" Read more Hmmmm… All of this is really valuable crowd-sourced input to the development effort. Hopefully Tesla is paying attention. These are normal situations in which the system can simply slow down, stop and relax. The objective is not "to get places as fast as possible, to hell with everything else". Chill! Alain
F. Lambert, May 23, "Tesla hacker green has brought to light a good example of why it’s important. He shared a picture of a Model 3 owner using Navigate on Autopilot and receiving a lane change suggestion into upcoming traffic:..
s shown on the image, the Model 3 is driving on a two-lane two-way undivided road and yet, Tesla’s Navigate on Autopilot is suggesting to move to the left lane even though there’s no vehicle to pass and there’s oncoming traffic in the left lane:..
I myself have a lot of experience with driving using Navigate on Autopilot in Quebec. I did notice that it would work on roads other than highways, including undivided roads, but I don’t remember it ever suggesting a lane change on those roads.
The reason behind the suggestion is unclear. It’s possible that the route required a left turn next and Autopilot believed that the car needed to be in the left lane to take a left turn. In this case, it would be a problem with Autopilot’s maps, but it’s not 100% clear.
It’s certainly worrying and serves as a great reminder to always stay vigilant and be ready to take control when using Autopilot or Navigate on Autopilot." Read more Hmmmm… I agree, It is almost definitely a problem with the map database (not an HD centimeter accuracy problem, just a most basic, how many lanes problem), given that there is no slow traffic ahead to pass, there is no reason for recommending a lane change.
1. the Tesla system is able to "see" sufficiently down the merge-to lane to make sure that there a clear portal/Volume ahead to enable passing. The concept of a clear "volume" ahead is VERY important; not just a clear surface ahead. A car needs more than a clear surface, it needs a clear volume such that the car can pass between and under any obstruction that may exist above the surface ahead. All too many, if not all, AV systems identify drivable surfaces ahead rather than drivable volumes. In most cases these are one and the same; however, when dealing with stationary objects ahead, especially, elevated stationary objects, these systems fail. They fail to reliably identify the height of those objects. Radars "see under" overpasses, overhead signs, tree canopies and the bottoms of trailers, and stopped fire trucks. Yipes!!!,
2. the Tesla system is smart enough to see that the time to collision with the oncoming truck in the passing lane is so small that there is no time to pass anybody, so forget about the lane change, and
3. the Tesla system realizes that since there is nothing ahead to pass then just stay in your lane and be happy. Again, this isn’t "rocket science" (How about SpaceX), see also Starlink and stuffin.space.) Alain
Half-baked stuff that probably doesn’t deserve your time
A. Li, May 24, "Late last year, Waymo launched the first autonomous ride service in Phoenix, Arizona. Before a wider launch, the Alphabet company has the big task of making self-driving cars inviting to the public. Waymo is now launching a contest that includes an all-expenses-paid trip to Phoenix. This “Meet Waymo contest” was announced on Friday, and invites the public “to meet the Waymo Driver during an exclusive weekend.” The “Waymo Driver” is how the company has recently referred to its self-driving vehicle.
We’ll fly you and a friend to Phoenix for a behind-the-scenes tour of our Operations Center and a ride in a Waymo self-driving car. For a chance to win, post a brief video telling us why you’re excited about self-driving technology and how you think it will change the world…" Read more Hmmmm… They must be kidding! Or is 9t05 Google’s Onion? I’m flabbergasted. Someone please explain. For what purpose?? They have nothing better to do? Don’t they know why they are spending a Billion a year trying to do this??? Alain
F. Lambert, May 22, "Elon Musk’s Boring Company has released a video of a race between two Tesla Model 3 vehicles: one inside the company’s tunnel under Los Angeles and the other traveling the same distance but on the surface streets. The company has a mile-long tunnel under Hawthorne near SpaceX’s headquarters in Los Angeles… " Read more Hmmmm… The video is comical. Its a mile long, why not walk? Alain
C’mon Man! (These folks didn’t get/read the memo)
Simply Click Bait
Calendar of Upcoming Events:
evening May 19 through May 21, 2019