http://smartdrivingcar.com/5.21-NTSB_TeslaDocket-062517
21st edition of the 5th year of SmartDrivingCars

Sunday, June 25, 2017

NTSB Opens Docket on Tesla Crash

Press release, June 19, "As part of its ongoing investigation into the fatal 2016 highway crash involving a Tesla Model S and a tractor-semitrailer truck near Williston, Florida, the National Transportation Safety Board on Monday opened the accident docket, releasing more than 500 pages of information.

System performance data downloaded from the Tesla revealed that the driver was operating the car using automated vehicle control systems: Traffic-Aware Cruise Control and Autosteer lane keeping systems.

The docket includes reports that cover various aspects of the investigation, including highway design, vehicle performance, human performance, and motor carrier factors. The crash reconstruction report, also included in the docket, provides a description of the crash sequence. The docket also includes interview transcripts and summaries, photographs, and other investigative material.  The docket contains only factual information collected by NTSB investigators; it does not provide analysis, findings, recommendations, or probable cause determinations. No conclusions about how or why the crash occurred should be drawn from the docket. Analysis, findings, recommendations, and probable cause determinations related to the crash will be issued by the Board at a later date.

The docket material is available at: https://go.usa.gov/xNvaE" Read more  Hmmmm… A few comments…
1.  Since lateral control (swerving) couldn’t have avoided this crash (the truck is almost 70 ft long (6 lanes wide) stretching broadside across the highway) , it doesn’t matter if Josh Brown ever had his hands on the steering wheel. That’s totally irrelevant. 
2.  Why didn’t autobrake kick in when the tractor part of the tractor-trailer passed in front of the Tesla?
3.  How fast was the truck going when it cut off the Tesla.  I couldn’t find the answer in 500 pages.   
4.  With sight distances of greater than 1,000 feet, why didn’t the truck driver see the Tesla?  Was it the drugs?
5.  This intersection invites "left-turn run-throughs" (no stop or yield and a 53 foot median and turn lane need to be crossed before one slips through a gap in two traffic lanes.  So you certainly roll into it, (plenty of room to stop if you see something coming) and if you don’t see anything, you hit it.  If you’re in the Tesla, you think you’ve been clearly seem, you expect the truck to stop, it doesn’t, you can’t believe it, BAM!  All in probably a second or so.
6.  The head injury description (Table 1 p2 of 3) certainly suggests that Joshua Brown was seated upright facing forward at impact.  The bilateral lacerations on the lower arm from the elbow to the wrist may indicate that he saw it coming in the last second and raised his arms in an attempt to protect his head.   The evidence reported doesn’t seem to suggest he saw this early enough to bend toward the passenger seat and try to pass underneath. 
7.  About 40 feet of tractor and trailer passed directly in front of the Tesla prior to impact.   Depending on how fast the truck was traveling, that takes some time.  Has NTSB run Virtual Reality simulations of various truck turn trajectories and analyzed what the truck driver and the Tesla driver could/should have seen?  Seems like a relatively simple thing to do.  We know what the Tesla was doing prior to the crash (going 74 mph straight down the road.) and we know where it hit the truck.  How fast the truck was traveling doesn’t seem to be known.
8. Why wasn’t there any video captured from the Tesla.  Didn’t that version of the MobilEye system store the video; I guess not, 🙁 
Anyway, lots to read in the 500 pages, but there is also a lot missing.  I’m not linking the many articles reporting on this because I disagree with many of their interpretations of the facts reported by NTSB.   Please reach your own conclusions.   Alain

Tesla, others seek ways to ensure drivers keep their hands on the wheel

D. Shepardson,  June 23, "…In a report this week on the May 2016 crash of a Tesla Inc (TSLA.O) Model S that killed a driver who was using Autopilot, the National Transportation Safety Board demonstrated that users could mostly keep their hands off the wheel for extended periods despite repeated warnings from the vehicle…." Read more  Hmmmm… So what???  The Tesla didn’t steer off the road.  It didn’t bake even though 40 feet of tractor trailer passed directly in front of it.  Even if you have a hand on the steering wheel, it doesn’t mean that you are paying attention.   If the automated emergency braking system worked properly it may have seen the cross traffic coming and applied the brakes.  Let’s make that stuff work!  Alain

CONGRESS FINALLY GETS GOING ON THAT REGULATING ROBOCARS THING

A. Marshall, June 21, "SEVEN YEARS AFTER Google started developing robocars, 13 months after a Florida man died in a Tesla Model S that was driving itself, and almost a year after self-driving Ubers started picking up passengers in Pennsylvania, Congress might actually start regulating autonomous vehicles….
No longer. Maybe. Last week, the Senate published bipartisan principles outlining what the legislation might look like. House Republicans, meanwhile, started circulating the drafts of a 14-bill package making it easier for federal regulators to make all the rules. …"  Read more  Hmmmm… Way too early!  We are still at the very beginning. If the Benz Patent on Jan 19, 1886 is the beginning of the transition from the buggy whip to the steering wheel.  Little got off the ground until 22 years later when Henry Ford unveiled his Model T on Oct 1, 1908.  If we take the 2005 DARPA Challenge as equivalent to the Benz patent and we really don’t have a Henry Ford in sight, then we are probably at "1898" in the evolution from the Steering wheel to the Chauffeured Driverless "Dowton Abbey" Mobility-on-Demand.  Can you imagine Congress in 1898 regulating the car????  Henry wouldn’t have had a chance.  So, today, please No!  Way too early.  Give the technology a chance before you go in there and screw it up.  Alain

Arizona getting ahead of autonomous vehicle industry by stepping aside

R. Randazzo, June 23. "With major testing by Waymo, Uber, General Motors, Ford and Intel, Arizona is more than holding its own in the race to attract the self-driving car industry.

Though 22 states have either passed legislation or executive orders addressing the industry,  Arizona has several things working in its favor….
…A 2015 executive order from Gov. Doug Ducey didn’t hurt, either, he said. The order aimed to make suggestions if needed, but not over-regulate the fledgling industry. Ducey directed a committee to advise the Department of Transportation, law enforcement and universities on how they can advance the deployment of the vehicles on state roads….
….Arizona’s oversight group has met just twice in the last year, and found no reason to suggest any new rules or restrictions on autonomous vehicles, so long as they follow traffic laws. The group found no need to suggest legislation to help the deployment…."  Read more  Hmmmm… Nice!  Alain 

Ready or not, states get ready for self-driving revolution

M. della Cava, June 25, "Self-driving vehicles are synonymous with sophisticated sensors producing terabytes of data being analyzed by powerful computers. But it seems the success of this transportation revolution hinges on a decidedly low-tech material: Paint.

That’s because when it comes to getting the nation’s infrastructure ready for autonomous traffic, the most critical upgrade amounts to making sure the lines on our 4 million miles of roads are solid, bright and preferably white so they can be picked up by computer vision gear…"  Read more  Hmmmm… And they can easily be seen by us as we continue to drive.  Alain 

AUTONOMOUS VEHICLES | SELF-DRIVING VEHICLES ENACTED LEGISLATION

Blog, June 19 "…NCSL has a NEW autonomous vehicles legislative database, providing up-to-date, real-time information about state autonomous vehicle legislation that has been introduced in the 50 states and the District of Columbia…"Read more  Hmmmm… Excellent source of enacted autonomous vehicle legislation.  Alain 

   Inside Travis Kalanick’s Resignation as Uber’s C.E.O.

M. Issac, Jume 21, "At a Chicago hotel on Tuesday, two venture capitalists presented Mr. Kalanick with a list of demands, including his resignation by the end of the day…"   Read more  Hmmmm… Those with the money, call the shots. Money talks, Brashness walks.  Always been that way.  Alain

The End of Car Ownership

T. Higgins, June 20, "Cars are going to undergo a lot of changes in the coming years.  One of the biggest: You probably won’t own one.

Thanks to ride sharing and the looming introduction of self-driving vehicles, the entire model of car ownership is being upended—and very soon may not look anything like it has for the past century…" Read more  Hmmmm… WSJ weighing in without anything new.  I wish that they would not confuse self-driving (which will reinforce the current car ownership model) with driverless, (which enables affordable mobility as a service and severely deminishes car ownership).  Just as some people still own horses, some people will still own their own car.  They just wont use it much..  Alain

Remote operation of vehicles with 5G

Mobility report, June 2017, "In the near future it will be a common occurrence to see driverless buses on city streets. A key step towards introducing autonomously driven buses into the public transport system is the development of remote monitoring and control capabilities, which will help to ensure safety, the Ericsson Mobility Report describes…." Read more  Hmmmm… Seems very likely that remote monitoring with possible operation may well be an evolutionary step to driverless.   I suspect that when Google sent their driverless bug around without test engineers that remotely monitored the operation in an intense fashion.  Alain

Bosch to Build $1.1 Billion Chip Plant for Self-Driving Cars

E. Behrmann, June 19, "Auto supplier Robert Bosch GmbH will build a 1 billion-euro ($1.1 billion) semiconductor plant, the biggest single investment in its history, as the maker of brakes and engines prepares for a surge in demand for components used in self-driving vehicles…" Read more  Hmmmm… Very interesting.  Bosch is protecting its "make or buy" options by making its own.  Seems like a stretch with this technology.  What edge do they have? Alain

  Uber Made an Unusual Legal Pledge to the Head of Its Driverless Car Project

J. Rosenblatt, June 23, "Uber made an unusual commitment to the engineer it hired to lead its driverless car project: It would cover the costs of legal actions against him over information stored in his head from his previous job at Waymo.   That promise — buried in the fine print of an otherwise straightforward employment contract for an executive — emerged in documents unsealed last week in San Francisco federal court. …
The Alphabet Inc. unit’s claims were bolstered Wednesday when it told the court Uber has said that Levandowski informed then-Chief Executive Officer Travis Kalanick more than a year ago that he had five discs containing Google data. Kalanick told him not to bring the information with him to Uber, and Levandowski said he then destroyed the files, according to the filing…." Read more  Hmmmm… Very interesting.  Alain

  Tesla loses 3 more computer vision experts on the Autopilot team

F. lambert, June 23, "After several changes in the Autopilot team’s leadership over the past few months, Tesla ended up hiring one of the leading AI and computer vision researchers this week.

At the same time however, we learn that several of the team’s top computer vision experts have left the company over the past few weeks. Read more  Hmmmm…All of this is VERY competitive with very limited supply. Also see  Tesla Autopilot: early test shows improvements w/ latest ‘silky smooth’ control algorithm and cyclist detection Alain

Uber’s Other Big Problem: Driverless Cars Aren’t Ready Yet

P. Burrows, June 22, "Its new CEO will inherit many problems, but a business plan based on the elusive dream of driverless cars is the largest…." Read more  Hmmmm…Pretty obvious.  Alain


Some other thoughts that deserve your attention

A Street Fight Among Grocers to Deliver Your Milk, Eggs, Bananas

 R. Abrams, June 24, "Every couple of days, Sinclair Browne fights through traffic in Times Square, squeezes his delivery truck into a parking spot, walks up four flights of stairs and delivers groceries to a guy whose order he knows by heart.

“I’m fast,” said Mr. Browne, slicing his hands in the air, ninja style. “In and out, in and out.”  Delivering food requires military precision: Bananas can’t get cold. Produce can’t get warm. Eggs, of course, must not get broken. And people expect their food to arrive at specific times.

Mr. Browne, 40, is a driver for the online grocery business Peapod. He plays the most important role in solving the biggest problem vexing the online grocery industry: moving food, undamaged and unspoiled, from the warehouse to the customer’s house. It’s known as the last-mile problem…." Read more  Hmmmm… If Amazon is really going to disrupt the Grocery business with Driverless vans (Hopefully) or drones (Unlikely) they’ll need to have paid attention to this article.  Alain

  Easiest Path to Riches on the Web? An Initial Coin Offering

N. Pooper, June 23, "A new crop of technology entrepreneurs is forgoing the usual routes to raising money. The entrepreneurs are not pitching venture capitalists, selling stock in an initial public offering or using crowdfunding sites like Kickstarter.

Instead, before they even have a working product, they are creating their own digital currencies and selling so-called coins on the web, sometimes raising tens of millions of dollars in a matter of minutes.

The pitch is that once the products are up and running, the currencies — with names like BAT, Mysterium and Siacoin — will be redeemable for services like data storage or anonymous internet access, and could appreciate in value in the meantime….

Known as initial coin offerings,…"   Read more  Hmmmm… These could be new transit tokens for those investing in Driverless cars.  ???  Alain

A discussion about AI’s conflicts and challenges

N. Lomas, June 17, Thirty five years ago having a PhD in computer vision was considered the height of unfashion, as artificial intelligence languished at the bottom of the trough of disillusionment.

Back then it could take a day for a computer vision algorithm to process a single image. How times change…." Read more    Hmmmm… Yup!, except…"Blake sees cracking unsupervised learning as the next big challenge for AI researchers to grapple with…." I disagree.  Unsupervised learning is an oxymoron.  There can be no learning unless there is an arbiter that states/reinforces truth/untruths, right/wrong, good/bad, …"Labeled data"  very valuable. "un-labeled data" is worthless until it gets labeled correctly.  Incorrectly labeled data is worse than worthless.  Seems all very simple and straightforward to me.  Alain

  Uncle Sam Wants Your Deep Neural Networks

C. Metz, June 22, "The Department of Homeland Security is turning to data scientists to improve screening techniques at airports.

On Thursday, the department, working with Google, introduced a $1.5 million contest to build computer algorithms that can automatically identify concealed items in images captured by checkpoint body scanners.

The government is putting up the money, and the six-month contest will be run by Kaggle, a site that hosts more than a million data scientists that was recently acquired by Google.

Although data scientists can apply any technique in building these algorithms, the contest is a way of capitalizing on the progress in a technology called deep neural networks,…" Read more    Hmmmm… This is all about the training data.  If DHS has invested in capturing the scans and their associated operator interpretation of those scans ("Labeled Data"), then the problem is trivial. If they haven’t (which it seems as if they haven’t), then its extremely challenging, if not impossible.  As is any of these problems, its all about the labeled training data.  Alain

Can NVIDIA’s Strong Growth Momentum Continue in Fiscal 2018?

P. Tanner, June 15, "As we move toward a data-centric world, the demand for computing power is growing. The technology industry (QQQ) is transitioning from microprocessor computing to GPU (graphics processing unit) accelerated computing as Moore’s law slows.

Moore’s law states that the number of transistors in a microprocessor would double every two years, improving its performance and power efficiency and reducing the cost of manufacturing….NVIDIA (NVDA) is at the epicenter of the GPU-accelerated computing trend and has seen strong growth momentum for more than a year. The company is in the early stage of the artificial intelligence and self-driving cars supercycle. This is visible from NVIDIA’s 60% and 70% CAGR1 in its Data Center and Automotive segments, respectively, between fiscal 2014–2017." …"   Read more    Hmmmm… Interesting.  Alain


On the More Technical Side

http://orfe.princeton.edu/~alaink/SmartDrivingCars/Papers/



Half-baked stuff that probably doesn’t deserve your time


C’mon Man!  (These folks didn’t get/read the memo


Calendar of Upcoming Events:


July 10 -14, 2017
Hilton San Francisco Union Square



2nd Annual Princeton SmartDrivingCar Summit
May 16 & 17, 2018
Princeton University
Princeton, NJ
Save the Date


Recent Highlights of:

Monday, June 19, 2017

Amazon Deal for Whole Foods Starts a Supermarket War

R. Abrams, June 16, "Shares of Walmart, Target, Kroger and Costco, the largest grocery retailers, all tumbled on Friday. And no wonder..  Grocery stores have spent the last several years fighting against online and overseas entrants. But now, with its $13.4 billion purchase of Whole Foods, Amazon has effectively started a supermarket war.  Armed with giant warehouses, shopper data, the latest technology and nearly endless funds — and now with Whole Foods’ hundreds of physical stores — Amazon is poised to reshape an $800 billion grocery market that is already undergoing many changes…."  Read more  Hmmmm… Since Jeff Bezos doesn’t need to have you impulse buy on your walk through the store while you get a quart of milk, he simply has to get you click on organic milk and he’ll present you with everything you absolutely can’t checkout without.  All he then needs is to get all those impulse buys (and the quart of organic milk) to your home from the hundreds of physical stores.  That’s where low speed driverless local delivery vans come in (operating initially in the early morning hours when the streets connecting those stores to our houses are completely empty and simply drop off everything you’ll need for the day ahead in your "Amazon Box" that’s replaced your 20th Century mailbox).  So in the end it will be Jeff Bezos’86 battling Eric Schmit’76 for deploying the first fleets of driverless vehicles sharing our neighborhood streets. If they should decide to join forces and have these vehicles providing mobility whenever anyone wants to travel and moving groceries and other goods the rest of the time, watch-out!!! Then everybody wins!! (except Walmart, Target, Kroger and Costco)  See also..Amazon and Whole Foods and Self-Driving Cars    Alain

Monday, June 12, 2017

G.M. Wants to Drive the Future of Cars That Drive Themselves

B. Vlasic, June 4, "… How would it react, for example, when it reached an intersection as a light turned yellow?  Driving in a situation like that, “you have to make a decision,” she recalled in a recent interview. “Generally if you decide to go, you decide to speed up. Or you stop.” If the technology works, she said, it will make the right decision: “The car knows.”

In beginning (sic) to assemble fully automated Bolts in January, G.M. was a step ahead of Google and Uber, which are converting mass-market minivans and sedans into driverless models. It went beyond what Tesla has achieved with autonomous controls on its own models. And it reflected the feverish competition underway…" Read more  Hmmmm… C’mon Bill, GM might be a step ahead of Google (who is not in this anymore) , but it is nowhere near Waymo.  And what is a "fully automated" Bolt?  It is certainly NOT "fully" anything.   Seems like the same-old GM but now with new ‘Corinthian Leather’ to sell to consumers through their network of dealers.  That leather will boost sales in the short run, and decrease the carnage on our roadways , which is great while keeping everyone in their same-old comfort zone of owning their own.   Unfortunately, the Bolt’s  "Corinthian Leather" turns to Flea & Moth Infested Burlap during parts of every trip, thus requiring a human driver to be ready to come to the rescue.  Consequently, the Bolt does NOT deliver the elevator-like mobility that some/many envision when someone says "cars that drive themselves" without explicitly adding in bold "only some of the time.  It won’t be able to provide mobility to the young, those under the influence, those that don’t have a driver’s license,… The Bolt is of no real help to Lyft or Uber or Didi or …but many people will buy Bolts!   Alain

Sunday, May 28, 2017

Rethinking Mobility: The ‘pay-as-you-go’ ca: Ride hailing, just the start

 S. Burgstaller, May 23,"The c.$7 tn global mobility market is speeding into the era of the “pay-as-you-go” car. Ride-hailing services such as Uber and Didi are pioneering a ‘cloud’ mobility system, which is using data to change
how the wealthiest cities move. In Rethinking Mobility, we model how the ride-hailing opportunity can grow to $285 bn by 2030, and is the precursor to a broader technological and social transformation.  We examine how the market might live up to the high valuations of its pioneers, why car sales may prove surprisingly resilient despite the change, and where automakers have a chance to transform their profitability as operators of fleets of autonomous cars…." Read more Hmmmm… Nice to see GoldMine Sachs finally weigh in.  The report is chock full of information and there is a lot here to absorb. 

The big impact will be if we ever get to Driverless without which you don’t replace even one Uber driver.  
Without Driverless, the issue centers on the 8x penetration of hailing rides.  At 8x  only car rental and little else is effected.  At 80x it effects car ownership but there will not be enough gig workers to support it.  So it doesn’t scale without Driverless.

With Driverless, then it is all about ridesharing as with elevators.  If it is as easy as elevators, then car ownership  diminishes greatly.

The report doesn’t respect the enormous difference between Driverless and Non-driversless (Self-driving and Safe-Driving; Levels 0 -> 4).  It seems to  assumes Driverless, yet it does not deal with the likelihood that Driverless will be achieved and fails to realize/identify the enormous forces that may come to bear that will attempt to derail Driverless at all costs.  The strongest of which may well be the "GMs" of this world.  GMs are all about Self-driving which REQUIRES a driver ( thus consumer ownership) and perpetuates their 100 year old business model.  Driverless scales ‘cloud mobility’ beyond the ‘8x’ limits of a gig economy and enables horizontal ‘cloud mobility’ to become as ubiquitous as the elevator is in vertical mobility.  Yes, there are still stair cases, and private ‘cloud-mobility" elevators for the 0.01%, but the masses will just grin&share the on-demand ‘cloud-mobility’ elevators without a 2nd thought.  Driverless assuaged vertical mobility anxiety. Driverless is the critical technological element that will assuage horizontal mobility anxiety and enable widespread horizontal ‘cloud mobility’.

Communities may find, as tall buildings have found, that they really work best (even at all) if they accommodate shared ‘cloud’ mobility and provide it for ‘free’ simply because it is so effective in capturing the enhanced land values that are unlocked by such mobility.  We’ve always been able to walk up and down a couple of flights of stair, but once we were easily able to go (via on-demand ‘cloud’ mobility available 24x7x365) more than four or so, then the sky became the limit. Are similar horizontal land values waiting to be unlocked if they simply pick up the tab for that on-demand horizontal ‘cloud’ mobility?  If so, then the GMs of this world are in a heap of trouble.  Alain

Tuesday, May 23, 2017

Princeton SmartDrivingCar Summit

May 18, Enormously successful inaugural Summit starting with the Adam Jonas video and finishing with Fred Fishkin’s live interview with Wm. C Ford III.  In between, serious engagement among over 150 leaders from Communities at the bleeding edge of deployment, Insurance struggling with how to properly promote the adoption of technology that may well force them to re-invent themselves and AI (Artificial Intelligence) and the various technologies that are rapidly advancing so that we can actually deliver the safety, environmental, mobility and quality of life opportunities envisioned by these “Ultimate Shared-Riding Machines”.

Save the Date for the 2nd Annual… May 16 & 17, 2018, Princeton NJ  Read Inaugural Program with links to Slides. Fishkin Interview of Summit Summary and Interview of Yann LeCun Read Inaugural Program with links to Slides. Hmmmm… Enormous thank you to all who participated.  Well done!  Alain

Sunday, May 14, 2017

Exploring the Bear Case: Distracted Driving + ADAS = $7 Trillion of Used Values at Risk

A. Jonas, Feb 1 "A sharp rise in traffic death & rapid growth of semiautonomous tech as standard equipment can accelerate the obsolescence of used cars, with potentially negative implications for secondhand values, auto credit & SAAR. We see elevated auto credit risk & avoid used car exposure….
…One could reasonably argue that if a technology can save 10k or 20k lives and hundreds of thousands of injuries per year in the US it should be (1) affordable and (2) not be optional equipment. Contrary to this, we found the majority of models currently available either do not offer active safety features or offer them only as optional equipment at prohibitively high costs. Our key takeaways are summarized below:…" Read More Hmmmm… First, sorry that I just saw this excellent report.  On top of the enormous substance, this report doesn’t mention that some/many of these systems don’t work as well as they should.  Some don’t brake if the the object ahead is stationary, others get confused with white back-lighting, others only apply the brake after the driver starts applying the brake and others only apply the brakes up to a 50% level.  Here we are trying to let drivers take hands of wheels and feet off pedals, yet we don’t have Safe-driving Cars that actually work (…experiencing essentially no false positives or false negatives) .  Alain

Monday, May 8, 2017

Walt Disney World plans to deploy driverless shuttles in Florida

R. Mitchell, Apr 28, "Walt Disney World in Florida appears poised to launch the highest-profile commercial deployment of driverless passenger vehicles to date, testing a fleet of driverless shuttles that could cart passengers through parking lots and around its theme parks.

According to sources with direct knowledge of Disney’s plans, the company is in late-stage negotiation with at least two manufacturers of autonomous shuttles – Local Motors, based in Phoenix, and Navya, based in Paris. It’s unclear whether contracts would go to both or just one of the companies…." Read More Hmmmm…This is exciting and substantial especially if it will be justified purely on its ability to deliver mobility, not entertainment, and will be financially self-sufficient.  Since it will be operating on Disney property, Disney can pretty much do as Disney wishes without having to be burdened by regulation meant to alleviate anxiety about the new and unfamiliar.  This is really exciting! Alain

Thursday, April 27, 2017

Waymo to Offer Phoenix Area Access to Self-Driving Cars

D. Streitfeld, Apr, 25, "…On Tuesday, the company was to announce the next phase of testing: putting ordinary people inside its Chrysler minivans and Lexuses….Only those who live in Chandler, Mesa, Tempe and Gilbert — roughly the southeastern Phoenix area — will be eligible for the program. And the cars, for that matter, will not take them anywhere else — no weekend jaunts to the Grand Canyon. Read More Hmmmm... Here we go!  Very conservative, but the path ahead is clear.  In 2013 they said that they were going to do this in 2017!   This is the beginning of real commercialization. Congratulations!  This is a major milestone.  Alain

Tuesday, April 17, 2017

Don’t Worry, Driverless Cars Are Learning From Grand Theft Auto

D. Hall, Apr 17, "In the race to the autonomous revolution, developers have realized there aren’t enough hours in a day to clock the real-world miles needed to teach cars how to drive themselves. Which is why Grand Theft Auto V is in the mix.
The blockbuster video game is one of the simulation platforms researchers and engineers increasingly rely on to test and train the machines being primed to take control of the family sedan. Companies from Ford Motor Co. to Alphabet Inc.’s Waymo may boast about putting no-hands models on the market in three years, but there’s a lot still to learn about drilling algorithms in how to respond when, say, a mattress falls off a truck on the freeway….The idea isn’t that the highways and byways of the fictional city of Los Santos would ever be a substitute for bona fide asphalt. But the game “is the richest virtual environment that we could extract data from,” said Alain Kornhauser…"  Read More Hmmmm... Well…we have a slightly different view of history wrt to GTA5.  The ‘Alain view’ is that Chenyi Chen*16 independently started investigating the use of virtual environments as a source of Image – Affordances data sets to use as the training sets in a ‘Direct Perception’ approach to creating a self-driving algorithm.  Images of the road ahead are converted into the instantaneous geometry that is implied by those image.  An optimal controller then determines the the steering, brake and throttle values to best drive the car.  The critical element in that process are the Image – Affordances data sets which need to be pristine.  Chenyi demonstrated in his PhD dissertation , summarized in the ICCV2015 paper,  that by using the pristine Image – Affordances data sets from an open-source game TORCS one could have a virtual car drive a virtual race course without crashing.  More importantly, when tested on images from real driving situations, the computed affordances were close to correct.

This encouraged us to look for more appropriate virtual environments. For many reasons, including: "wouldn’t it be amazing if ‘Grand Theft Auto 5’ actually generated some positive ‘redeeming social value’ by contributing to the development of algorithms that actually made cars safer; saving grief, injuries and lives".  Consequently, in the Fall of 2015, Artur Filipowicz’17 began to investigate using GTA5 to train Convolutional Neural Networks to perform some of the Direct Perception aspects of automated driving.  With Jeremiah Liu, he continued his efforts in this direction last summer which were presented at TRB in January.  Yesterday, he and Nyan Bhat’17 turned in their Senior Theses focused on this topic.

Indeed, GTA5 is a rich virtual environment that begins to efficiently and effective address the data needs of Deep Learning approaches to safe driving.    Alain

Monday, March 20, 2017

Uber’s autonomous cars drove 20,354 miles and had to be taken over at every mile, according to documents

J. Bhuiyan, Mar 16, "Some of Uber’s self-driving cars aren’t driving as smoothly as the company hoped they would. Documents circulated throughout the company’s self-driving group, which Recode obtained, gives us a first look at the progress of the ride-hail company’s robot cars in Pennsylvania, Arizona and California.
The top line: Uber’s robot cars are steadily increasing the number of miles driven autonomously. But the figures on rider experience — defined as a combination of how many times drivers have to take over and how smoothly the car drives — are still showing little progress….
For example: During the week ending March 8, the 43 active cars on the road only drove an average of close to 0.8 miles before the safety driver had to take over for one reason or another…
The good news is the number of miles between these “critical” interventions has recently improved. Last week, the company’s cars drove an average of approximately 200 miles between those types of incidents that required a driver to take over…" Read more Hmmm… Waymo is so incredibly far ahead.  Even with these statistics, it depends on when and where the miles were drive.  It is relatively unchallenging in some places at some times, especially if you’ve experienced it many times before. Its all about being able to handle the unexpected to achieve Driverless. Uber accrues no substantive value until it reaches Driverless.  Self-driving’s only value is as a way/process to achieve Driverless.  Alain

Friday, March 10, 2017

Robot cars — with no human driver — could hit California roads next year

R. Mitchell, Mar 10, "California is back on the map as a state that’s serious about welcoming driverless cars.Truly driverless cars — vehicles with no human behind the wheel, and perhaps no steering wheel at all — are headed toward California streets and highways starting in 2018…
The regulations lay out “a clear path for future deployment of autonomous vehicles” in California, said Bernard Soriano, deputy director at the Department of Motor Vehicles…." Read more Hmmm… Congratulations Bernard!  This is fantastic news on the road to providing high-quality mobility for all.  It squarely addresses the fundamental need to efficiently re-position vehicles so that they can get to even those who can’t drive.  This is a real turning point for automated vehicles from self-driving toys for the 1% to affordable, environmentally friendly mobility for everyone.  Alain

Friday, February 24, 2017

Alphabet’s Waymo Alleges Uber Stole Self-Driving Secrets

M. Bergen, Feb 23, "It took Alphabet Inc.’s Waymo seven years to design and build a laser-scanning system to guide its self-driving cars. Uber Technologies Inc. allegedly did it in nine months.
Waymo claims in a lawsuit filed Thursday that was possible because a former employee stole the designs and technology and started a new company….Anthony Levandowski, a former manager at Waymo, in December 2015 downloaded more than 14,000 proprietary and confidential files, including the lidar circuit board designs, according to the complaint. He also allegedly created a domain name for his new company and confided in some of his Waymo colleagues of plans to “replicate” its technology for a competitor…." Read more   Hmmm…This is very serious.  So unfortunate.  🙁   Alain

Friday, February 17, 2017

Motor Vehicle Deaths in 2016 Estimated to be Highest in Nine Years

Press release, Feb. 15, "NSC offers insight into what drivers are doing and calls for immediate implementation of proven, life-saving measures…

With the upward trend showing no sign of subsiding, NSC is calling for immediate implementation of life-saving measures that would set the nation on a road to zero deaths:…" Read more  Hmmm…"Automated Collision Avoidance" or anything having to do with ‘Safe-driving Cars‘ is not mentioned anywhere in the Press Release.  One of us is missing something very fundamental here!!  So depressing!!  🙁   Alain

Friday, January 27, 2017

Serving the Nation’s Personal Mobility Needs with the Casual Sharing of autonomousTaxis & Today’s Urban Rail, Amtrak and Air Transport Systems

A. Kornhauser, Jan 14, "Orf467F16 Final Project Symposium quantifying implications of such a Nation-wide mobility system on Average Vehicle Occupancy (AVO), energy, environment and congestion, including estimates of fleet size, needed empty vehicle repositioning, and ridership implications on existing rail transit systems (west, east, NYC) and Amtrak of a system that would efficiently and effectively perform their ‘1st mile’/’last-mile’ mobility needs. Read more  Hmmm… Now linked are 1st Drafts of the chapters and the powerPoint summaries of these elements.  Final Report should be available by early February.  The major finding is, nationwide there exists sufficient casual ridesharing potential that a well–managed  Nationwide Fleet of about 30M aTaxis (in conjunction with the existing air, Amtrak and Urban fixed-rail systems)  could serve the vehicular mobility needs of the whole nation with VMT 40% less than today’s automobiles while providing a Level-of-Service (LoS) largely equivalent and in many ways superior than is delivered by the personal automobile today.  Also interesting are the findings as to the substantial increased patronage opportunities available to Amtrak and each of the fixed rail transit systems around the country because the aTaxis solve the ‘1st and last mile’ problem.  While all of this is extremely good news, the challenging news is that since all of these fixed rail systems currently lose money on each passenger served, the additional patronage would likely mean that they’ll lose even more money in the future. 🙁  Alain 

ODI (Office of Defects Investigation) Findings on Tesla AEB & AutoPilot

(Above link should work) Jan 19, "… Summary: …     NHTSA’s examination did not identify any defects in the design or performance of the AEB or Autopilot  systems of the subject vehicles nor any incidents in which the systems did not perform as designed.  AEB systems used in the  automotive industry through MY 2016 are rear-end collision avoidance technologies that are not designed to reliably  perform in all crash modes, including crossing path collisions.  The Autopilot system is an Advanced Driver Assistance  System (ADAS) that requires the continual and full attention of the driver to monitor the traffic environment and be prepared to take action to avoid crashes.  Tesla’s design included a hands-on the steering wheel system for monitoring driver engagement…
…  ODI analyzed data from crashes of Tesla Model S and Model X vehicles involving airbag deployments that occurred while operating in, or within 15 seconds of transitioning from, Autopilot mode. Some crashes involved impacts from other vehicles striking the Tesla from various directions with little to no warning to the Tesla driver.  Other crashes involved scenarios known to be outside of the state-of-technology for current-generation Level 1 or 2 systems, such as cut-ins, cut-outs and crossing path collisions…. 
…The Florida fatal crash appears to have involved a period of extended distraction (at least 7 seconds)…"
.Hmmm… nothing else is written about this nor is a basis given for  the ‘at least 7 seconds’.  Possibly the most important information revealed in this summary is Figure 11, p11: "…  Figure 11 shows the rates calculated by ODI for airbag deployment crashes in the subject Tesla vehicles before and after Autosteer installation.  The data show that the Tesla vehicles crash rate dropped by almost 40 percent after Autosteer installation…
…A safety-related defect trend has not been identified at this time and further examination of this issue does not appear to be warranted.  Accordingly, this
investigation is closed. "  Read more  Hmmm… WOW!!! . Every word of this Finding is worth reading.  It basically exonerates Tesla, states that AEBs (Automated Emergency Braking) systems don’t really work and aren’t designed to work in some scenarios (straight crossing path (SCP) and left turn across path (LTAP), see p 2,3).  …which suggests, to me, that DoT/NHTSA should be placing substantial efforts on making these systems really work in more scenarios.  And… there is the solid data that ‘AutoSteer" reduced Tesla crashes by almost 40%!!! WOW!! Will Insurance now finally get on-board and lead?  Alai

Friday, September 23, 2016

Federal Automated Vehicles Policy: Accelerating the Next Revolution In Roadway Safety

September 2016, "Executive Summary…For DOT, the excitement around highly automated vehicles (HAVs) starts with safety.  (p5)

…The development of advanced automated vehicle safety technologies, including fully self-driving cars, may prove to be the greatest personal transportation revolution since the popularization of the personal automobile nearly a century ago. (p5)

…The benefits don’t stop with safety. Innovations have the potential to transform personal mobility and open doors to people and communities. (p5)

…The remarkable speed with which increasingly complex HAVs are evolving challenges DOT to take new approaches that ensure these technologies are safely introduced (i.e., do not introduce significant new safety risks), provide safety benefits today, and achieve their full safety potential in the future. (p6)  Hmmm…Fantastic statements and I appreciate that the fundamental basis and motivator is SAFETY.  We all have recognized safety as a necessary   condition that must be satisfied if this technology is to be successful.  (unfortunately it is not a sufficient condition, (in a pure math context)). This policy statement appropriately reaffirms this necessary condition.  Alain

"…we divide the task of facilitating the safe introduction and deployment (…defines “deployment” as the operation of an HAV by members of the public who are not the employees or agents of the designer, developer, or manufacturer of that HAV.) of HAVs into four sections:(p6) Hmmm…Perfect! Alain

"…1. Vehicle Performance Guidance for Automated Vehicles (p6)…"  Hmmm… 15 Points, more later. Alain

"…2. Model State Policy (p7)   The Model State Policy confirms that States retain their traditional responsibilities…but… The shared objective is to ensure the establishment of a consistent national framework rather than a patchwork of incompatible laws…" Hmmm… Well done.  Alain

"…3. NHTSA Current Regulatory Tools (p7) … This document provides instructions, practical guidance, and assistance to entities seeking to employ those tools. Furthermore, NHTSA has streamlined its review process and is committing to…"   Hmmm… Excellent. Alain

"…4. New Tools and Authorities (p7)…The speed with which HAVs are advancing, combined with the complexity and novelty of these innovations, threatens to outpace the Agency’s conventional regulatory processes and capabilities. This challenge requires DOT to examine whether the way DOT has addressed safety for the last 50 years should be expanded to realize the safety potential of automated vehicles over the next 50 years. Therefore, this section identifies potential new tools, authorities and regulatory structures that could aid the safe and appropriately expeditious deployment of new technologies by enabling the Agency to be more nimble and flexible (p8)…"  Hmmm… Yes. Alain

"…Note on “Levels of Automation”  There are multiple definitions for various levels of automation and for some time there has been need for standardization to aid clarity and consistency. Therefore, this Policy adopts the SAE International (SAE) definitions for levels of automation. )  Hmmm… I’m not sure this adds clarity because it does not deal directly with the difference between self-driving and driverless.  While it might be implied in level 4 and level 5 that these vehicles can proceed with no one in the vehicle, it is not stated explicitly.  That is unfortunate, because driverless freight delivery can’t be done without "driverless"; neither can mobility-on-demand be offered to the young, old, blind, inebriated, …without "driverless".  Vehicles can’t be "repositioned-empty" (which (I don’t mean to offend anyone) is the real value of a taxi driver today).  So autonomousTaxis are impossible.

Also, these levels do not address Automated Emergency Braking  (AEB) Systems and Automated Lane Keeping Systems which are the very first systems whose on-all-the-time performance must be perfected.   These are the Safety Foundation of HAV (Highly Automated vehicles).  I understand that the guidelines may assume that these systems are already perfect and that "20 manufacturer have committed" to have AEB on all new cars, but to date these systems really don’t work.  In 12 mph IIHS test, few stop before hitting the target, and, as we may have seen with the Florida Tesla crash, the Level 2/3 AutoPilot may not have failed, but, instead, it was the "Phantom Level 1" AEB that is supposed to be on all the time.  This is not acceptable.  These AEB systems MUST get infinitely better now.  It is a shame that AEBs were were not explicitly addressed in this document.

"…I. Vehicle Performance Guidance for Automated Vehicles (p11) A. Guidance: if a vehicle is compliant within the existing FMVSS regulatory framework and maintains a conventional vehicle design, there is currently no specific federal legal barrier to an HAV being offered for sale.(footnote 7)  However, manufacturers and other entities designing new automated vehicle systems
are subject to NHTSA’s defects, recall and enforcement authority. (footnote 8)   . and the "15 Cross-cutting Areas of Guidance" p17)

In sum this is a very good document and displays just how far DoT policy has come from promoting v2v, DSRC and centralized control, "connected",  focus to creating an environment focused on individual vehicles that responsibly take care of themselves.  Kudos to Secretary Foxx for this 180 degree policy turn focused on safety.   Once done correctly, the HAV will yield the early safety benefits that will stimulate continued improvements that, in turn, will yield the great mobility, environmental and quality-of-life benefits afforded by driverless mobility. 

What are not addressed are commercial trucking and buses/mass transit.  NHTSA is auto focused, so maybe FMCSA is preparing similar guidelines.  FTA (Federal Transit Administration) seems nowhere in sight.  Alain

Tuesday, July 5, 2016

May 7 Crash

Hmmm…What we know now (and don’t know):

Extracting Cognition out of Images for the Purpose of Autonomous Driving

Thursday, March 17, 2016

U.S. DOT and IIHS announce historic commitment of 20 automakers to make automatic emergency braking standard on new vehicles

Sunday, December 19, 2015

Adam Jonas’ View on Autonomous Cars

Video similar to part of Adam’s Luncheon talk @ 2015 Florida Automated Vehicle Symposium on Dec 1.  Hmmm … Watch Video  especially at the 13:12 mark.  Compelling; especially after the 60 Minutes segment above!  Also see his TipRanks.  Alain


This list is maintained by Alain Kornhauser and hosted by the Princeton University LISTSERV.

***************************************************************************************************************
This list is maintained by Alain Kornhauser and hosted by the Princeton University LISTSERV.

Unsubscribe | Re-subscribe

%d bloggers like this: