6.51-PlayItSafe-112918

https://www.princetondiary.com/smartdrivingcar/6.51-PlayItSafe-112918
51st edition of the 6th year of SmartDrivingCars

Thursday, November 22, 2018

Waymo’s Cars Play It Safer After Incidents and ‘Driver Fatigue’

A. Efrati, Nov 27, “Waymo has only weeks to meet its self-imposed deadline to launch a public taxi service using fully automated cars by the end of 2018. And right now, that deadline looks tough for the company to meet. The Information has learned that within the past month or so, due to concerns about safety, the Alphabet company put so-called “safety drivers” back behind the wheel of its most advanced prototypes, ending a year-long period in which those people generally sat in the passenger or back seat.

Meanwhile, The Information also has learned that Waymo is only testing its most advanced prototypes in about 60 square miles, or roughly 5% of the Phoenix metropolitan area, say people with knowledge of the situation….” Read more Hmmmm…. No problem. 5% is a very large area in which to start. And having attendants onboard is also OK, in the beginning. Not much would be saved or gained by removing them (except some machoism which has no real value.). It is the only way to go in the beginning because safety is fundamentally critical and much still needs to learned and improved. Once safety has been demonstrated in this “5%” the attendants can disappear and can move on to be attendants in the next 10%, and so on… This is the responsible market launch scenario. Alain

Smart Driving Cars Podcast Episode 69 – Chunka Mui

F. Fishkin, Nov 29, “What will it take for driverless vehicles to become a leading form of transportation? Futurist and author Chunka Mui joins Princeton’s Alain Kornhauser and co-host Fred Fishkin for Episode 69 of the Smart Driving Cars podcast. Plus…Waymo, GM, Amazon and more. Tune in and subscribe! ” Hmmmm…. Now you can just say “Alexa, play the Smart Driving Cars podcast!” . Ditto with Siri, and GooglePlay. Alain
Real information every week. Lively discussions with the people who are shaping the future of SmartDrivingCars. Want to become a sustaining sponsor and help us grow the SmartDrivingCars newsletter and podcast? Contact Alain Kornhauser at alaink@princeton.edu! Alain

15 Hurdles To The Industrialization Of Driverless Cars (Part 1 Of 3)

Chunka Mui, Nov 27, “Will the future of driverless cars rhyme with the history of the Segway? The Segway personal transporter was also predicted to revolutionize transportation. Steve Jobs gushed that cities would be redesigned around the device. John Doerr said it would be bigger than the Internet. The Segway worked technically but never lived up to its backers’ outsized hopes for market impact. Instead, the Segway was relegated to narrow market niches, like ferrying security guards, warehouse workers and sightseeing tours.

One could well imagine such a fate for driverless cars (a.k.a. AVs, for autonomous vehicles). The technology could work brilliantly and yet get relegated to narrow market niches, like predefined shuttle routes and slow-moving delivery drones. Some narrow applications, like interstate highway portions of long-haul trucking, could be extremely valuable but nowhere near the atmospheric potential imagined by many—include me, as I described, for example, in “Google’s Driverless Car is Worth Trillions“….” Read more Hmmmm… Excellent (but , of course, I have a dfew comments… Safety is paramount (in my view. I assume that Chunka either implies it in his “Trust”or sees it as so paramount that it is implied as part of his “technical feasibility” which overarches everything.

I see two worlds… one which is largely what we have now except there is more automation beyond an automatic transmission. This world really doesn’t changed much…. The 2nd world in which we have safe (must necessarily be, else DoA) Driverless Mobility/Transportation aaS. I look at Chunka’s 15 hurdles as this Driverless (very appropriately termed) world. I don’t see the need, nor desire (nor maybe even operational feasibility) for personal ownership of these anymore than I see a need/desire/feasibility to have a personal ownership of my own private elevator in my High-rise apartment building (unless I’m The Donald). Else, I love the15 hurdles. Alain

Tesla owners have driven 1 billion miles with Autopilot activated

F. Lambert, Nov 29, “Tesla’s global fleet recently reached half a million vehicles, and all of them added over the past three years have been equipped with Autopilot hardware, which has been gathering data for Tesla’s driver assist program.

The automaker now says that the Tesla owners have driven 1 billion miles with Autopilot activated as of today. Tesla says that it represents “10% of the total mileage driven by Tesla vehicles to date”, but it doesn’t mean that current Tesla drivers use Autopilot for 1 in 10 miles. It’s 10% of all the mileage of Tesla’s fleet, including vehicles that were not equipped with Autopilot hardware and those that have the hardware, but the owners didn’t buy Autopilot.

As we reported earlier this month, Tesla recently reached 10 billion electric miles with a global fleet of half a million cars. Now we’ve learned that 1 billion of those miles have been driven with the help of Autopilot….” Read more Hmmmm…. This is a totally impressive accomplishment and demonstrates that this Self-driving system does work and is used responsibly. I doubt that “conventional cruise control” is used 10% of the miles that cars so equipped are driven. (I know of no reporting of this statistic. I’m also unaware that any “roadway authority” has ever posted a sign encouraging users of their roadway to use conventional cruise control (even though use of such technology might actually smooth out traffic flow, increase capacity and save energy)).

All these data that Tesla has captured reveal which road segments, under which weather/roadSurface/congestion conditions, AutoPilot works really well, and where it is “challenged”. Will that information be made available to AutoPliot owners/users enabling them to use AutoPilot even more responsibly and to the “roadway authorities” so that they can improve the “paint/geometry” of their “challenged” road segments? Wouldn’t it be nice!!!

By the way, what are the equivalent statistics for”equivalent” systems from GM, Daimler, Subaru, Volvo, … ????? (or do these entities sell you their system and hope that you’ll never use it?? Each sold us conventional cruise control, but never seemed to have followed up to see if used it. Daimler has never asked me about my experiences with my “997 package”. Have they ever asked anyone?? Alain

The 5th Generation of Mobile Wireless is Arriving

M. Sena, Dec 2018, “ON OCTOBER 1ST, VERIZON, the largest mobile network oper-ator in the U.S., began offering installation of its 5G home broadband in Sacramento, CA with Houston, Indianapolis and Los Angeles to follow. Customers will be able to begin receiving service once the installation of small cells on util-ity poles and street lights is complete, with pricing set at $50 for VERIZON WIRELESS customers and $70 for non-cus-tomers. Those prices include all taxes and fees and do not require an annual contract. Customers who sign up for 5G home broadband service in the selected areas can expect bandwidth speeds of around 300Mbps, with peak speeds of nearly 1Gbps depending on location….

ON THE 3RD OF OCTOBER, 2018, the U.S. Department of Trans-portation issued Preparing for the Future of Transporta-tion: Automated Vehicles 3.0 (AV 3.0).5 A news release in-cluded the statement that “this document builds upon Au-tomated Driving Systems: A Vision for Safety 2.0 (ADS 2.0) (released on 12 September 2017 and reported on in the November 2017 issue of THE DISPATCHER) and expands the scope to provide a framework and multimodal ap-proach to the safe integration of AVs into the Nation’s broader surface transportation system.”… ” Read more Hmmmm… Another excellent Dispatcher. Thank you, Michael. Alain

AUTONOMOUS VEHICLES TECHNOLOGY REPORT

Woodside Technology Partners, Nov 2018, “Investment and M&A

In the first 9 months of 2018 there was a total $4.25B of venture capital invested across 87 companies in the AV (Autonomous Vehicle) sector, including ride sharing and EV companies. This compares to $2.1B of venture capital invested across 104 companies for the same period in 2017. There has been a total of $14.4B of venture capital invested into 183 AV companies over the last 3 years ….

Safety & Security

There is currently a vigorous debate within the automotive industry as too how safe is “safe enough” for autonomous vehicles. This debate has been fueled by recent AV/ADAS related fatalities – both Uber and Tesla cars have been involved in self-driving fatalities over the last 24 months. …

AI & Deep Learning

It is impossible to attend any of the numerous autonomous car conferences today without being struck by the extraordinarily complexity of autonomous car design. Humans can rely on “inference” and “common sense” to know what to do if they encounter a situation that they have never seen before (such as a private airplane that has had to make an emergency landing on the highway ahead). …

Sensors

Semiconductor AI Processors

Connected Cars

China… Read more Hmmmm… Interesting that there isn’t a segment on investments in the entities that will deliver the mobility of the ensemble of these investments in gizmos. Will those entities be successful enough to make the assembled components into a revolutionary mobility machine? Alain

Imagining What Happens When the Robots Take the Wheel… Book Review

M. DeBord, Nov. 27, “NO ONE AT THE WHEEL Driverless Cars and the Road of the Future By Samuel I. Schwartz 272 pp. PublicAffairs. $30.

An appalling statistic appears toward the end of “No One at the Wheel,” Samuel Schwartz’s valuable primer on self-driving cars: In the century since the automobile arrived on the scene, 70 million people have been killed by it, and four billion injured.

Schwartz, who served as New York City’s traffic commissioner in the 1980s, was nicknamed “Gridlock Sam” for his devotion to the conundrum of traffic (and for coining the loathsome term). He knows everything about how cars and people don’t get along, having been on the front lines. This book — written in an earnest, conversational style — is his attempt to grapple with a fresh threat that’s appeared after decades of progress….” Read more Hmmmm…. Well worth reading. Alain

Through All the Hype, Self-Driving Cars Remain Elusive

N. Mayersohn, Nov 27, “The relentlessly hyped arrival of autonomous vehicles looms as the greatest disruption in personal transportation since Henry Ford’s moving assembly line started producing Model T’s by the millions. One word in that statement — arrival — is, however, doing a disproportionate amount of work.

Self-driving vehicles, despite being the subject of breathless media reports and in automakers’ strategies, remain years from being available to private owners. Scores of companies hold permits to test autonomous cars in California, yet even leaders like Waymo, once Google’s self-driving project, are unwilling to commit to when such vehicles might be appearing in showrooms….” Read more Hmmmm…. 1st… Waymo’s Driverless won’t be appearing in any showroom in my lifetime and likely never. These are mobility machines that most anyone can use. They are totally unsuited for selfish personal ownership. The Self-driving versions, that by their fundamental nature require capable adult supervison, are personal/self mobility machines and are now widely available in showrooms (Volvo, Mercedez, Subaru, Tesla’s virtual showrooms, etc.) and work (see above). But these are two VERY different “animals” that should NOT be confused, just like a cat should NOT be confused with a tiger just because they are both animals. Or an amoeba and a human because they are free-living organisms. The press needs to stop suggesting that having tigers as pets or the admission to Princeton by amoebas “remains elusive”. Alain

Self-driving taxis are here. This is what it’s like to ride in one

G. Fowler, Nov 29, “… But in this Waymo, there’s actually still a human in the driver’s seat. Why’s that? That guy is our safety monitor. even though Waymo says it can run cars without humans inside, for most …all???… rides it’s still paying people to sit there and take over if needed. Sometimes, there’s an additional Waymo assistant in the car, too. …” Read more Hmmmm…. So in fact, Waymo is still where Uber was in Pittsburgh 2 years ago. Progress is being made, we have ignition, but we haven’t reached liftoff. Not likely to have a real liftoff in 2018. Really want to avoid… “obviously a major malfunction”… Alain

Autonomous vehicles: “Pedal to the metal or slamming on the brakes?” Worldwide regulation of autonomous vehicles

P. Keller, Nov 2018, “…Questions have been raised over whether the technology is “ready” to provide all of the benefits that have been touted for so long. Although the industry is trying to tackle some of these concerns, the role that lawmakers and local authorities will have in this process will and should be considerable. To be sure, there are only a few global, standard rules governing automobiles, and even fewer addressing autonomous vehicles. One of the few examples is the Vienna Convention on Road Traffic. Since 1968, this treaty has required that a human driver be in full control of and responsible for the behavior of the vehicle in traffic. This requirement, however, is now being revisited all over the world, as individual regions are tailoring their laws to reflect their own, local balance between safety and the development and use of self-driving vehicles.

The resulting panoply of rules could not be more varied: Read more Hmmmm…. Excellent status report of world wide regulation of AV. Alain

US Motor Vehicle Accident Deaths 1999-2016

J. Holman, Nov 2018, “This report examines United States motor vehicle accident (MVA) deaths during 1999-2016. The Society of Actuaries pursued the research as part of its ongoing longevity and mortality research initiatives. The objective of this research is to produce a long-term overview of these accident fatalities that may aid in a better understanding of them to support managing the selection and underwriting of insured risk. The report gives an overview of MVA deaths and analyzes them by person involved, age, gender, region, vehicle type, alcohol impairment, and by time series patterns. …

…The Centers for Disease Control (CDC) reports, per data in the Wide-ranging Online Data for Epidemiologic Research (WONDER), that in 2016 MVA deaths were 25% of all accidental deaths, which placed them second to poisonings as the highest accidental cause of death.”…

… Total annual deaths have declined from 41,717 in 1999 to 37,461 in 2016 (see Section 2.1 Figure 2). This is a 10.2% decrease that is equivalent to an annual 0.6% improvement (average geometric mean basis).” … disregard that “average geometric mean” because… “During 1999 to 2016, MVA deaths increased slightly to a peak of 43,510 in 2005, declined to a low of 32,479 in 2011, and since resumed an increasing trend to current levels. Most recently, MVA deaths increased by 5.6% in 2016 versus 2015. …

… The decrease of passenger deaths was much higher than that of drivers. This occurred because drivers carrying passengers had much lower fatal accident involvement rates, about five to ten per 100,000 drivers depending on age, than those driving alone. The subset of drivers carrying passengers experienced about the same rate of improvement as their passengers. ” … Sounds like another good reason to ride-share…. ….” Read more Hmmmm…. Very interesting, but given that over this period NHTSA’s crash mitigation initiatives became more commonplace. This adoption that likely resulted in saving the lives of many involved in severe crashes. A severe crash in 1999 involving a car without the advanced crash mitigation technology might have resulted in a fatality. That same crash in 2016 involving a car with advanced crash mitigation technology might not have resulted in a fatality, but instead a severe injury. What doesn’t seem to be known is the historical frequency and liability implications between 1999 and 2016 of the various types of crashes that resulted in deaths in 1999.

Why this is important is ..

if the crash frequencies are relatively constant, then the NHTSA crash mitigation initiatives haven’t saved many lives, or…
if the crash mitigation initiatives have saved lives, then the crash frequencies have been going up (drastically/alarmingly???). Why??? Alain

DÉVELOPPEMENT DES VÉHICULES AUTONOMES Orientations stratégiques pour l’action publique

A. Idrac, May 2018. “…Le développement du véhicule automatisé ou autonome représente un enjeu considérable pour les politiques des transports et de sécurité, notamment de sécurité routière, et l’industrie française. Plus précisément, les enjeux stratégiques identifiés par l’Etat sont de plusieurs ordres :… Read more Hmmmm…. France’s strategic plan for market penetration of automated vehicles. Alain

Black Box Data Reveals Pilots’ Struggle to Regain Control of Doomed Jet

J. Glanz, Nov 27, “Data from the jetliner that crashed into the Java Sea last month shows the pilots fought to save the plane almost from the moment it took off, as the Boeing 737’s nose was repeatedly forced down, apparently by an automatic system receiving incorrect sensor readings.

The information from the flight data recorder, contained in a preliminary report prepared by Indonesian crash investigators and scheduled to be released Wednesday, documents a fatal tug-of-war between man and machine, with the plane’s nose forced dangerously downward more than two dozen times during the 11-minute flight. The pilots managed to pull the nose back up over and over until finally losing control, leaving the plane, Lion Air Flight 610, to plummet into the ocean at 450 miles per hour, killing all 189 people on board….” Read more Hmmmm…. See also First Look at JT610 Flight Data. Here we have a situation in which there was redundancy… two angle of attack (AoA) sensors. Unfortunately, they didn’t agree, and it seems as if the “AI” had no way to tell which was correct, which was faulty or neither. Lesson… having redundant sensors without the ability to address what to do when they disagree is doomed to failure. What should have been done in this situation may well be as simple as… if they disagree, believe neither and revert to a situation that doesn’t require the sensor. (That doesn’t address the situation when both are equally inaccurate, but hopefully that is a very very rare occurrence.) (Also, the AI wasn’t smart enough to assess the data streams from the AoAs during the taxi phase before takeoff. If so, the AI would have noticed that the data streams disagreed, one giving values that are unrealistic while rolling along the ground and that the unrealistic one was substantially drifting during taxi. The AI should have made this known to the pilots prior to takeoff. In hindsight, vision is 20-20🙂,… but there is next time. 🙂 Alain

In China, your car could be talking to the government

E. Kinetz, Nov 29, “When Shan Junhua bought his white Tesla Model X, he knew it was a fast, beautiful car. What he didn’t know is that Tesla constantly sends information about the precise location of his car to the Chinese government.

Tesla is not alone. China has called upon all electric vehicle manufacturers in China to make the same kind of reports — potentially adding to the rich kit of surveillance tools available to the Chinese government as President Xi Jinping steps up the use of technology to track Chinese citizens….” Read more Hmmmm…. I’m so surprised!! Tell me it ain’t so. “…Chinese officials say the data is used for analytics to improve public safety, facilitate industrial development and infrastructure planning, and to prevent fraud in subsidy programs. …” Whew!!! Alain

G.M. to Idle Plants and Cut Thousands of Jobs as Sales Slow

N. Boudette, Nov 26, “General Motors will cut up to 14,000 workers in North America and put five plants up for possible closure as it abandons many of its car models and restructures to focus more on autonomous and electric vehicles, the automaker announced Monday.” Read more Hmmmm…. So the new jobs are in SmartDrivingCars ….See also… The existential threat facing the auto industry and Wall Street Reacts Favorably To GM Layoffs, Plant Closures This has nothing to do with SmartDrivingCars, if Waymo executes a successful lift-off, then ….” See also… GM axes Volt, Cruze, and Impala for North America in cost-cutting move See also… The existential threat facing the auto industry “… That’s because the way people get from Point A to Point B is about to undergo the most radical change since the early automobiles replaced horses. Autonomous driving vehicles are about to change how cars are driven, powered and used. In the future, far more people will probably buy rides, rather than cars… “We see these steps as necessary to ensuring the long-term sustainability and independence of GM,” wrote Adam Jonas, auto analyst for Morgan Stanley, in a note to clients…. .” T Alain

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

Amazon is releasing a $400 self-driving toy car that you can program yourself – and it’s launching a racing league to test your skills

R. Chan, Nov 28, “Amazon Web Services, the retail giant’s massively profitable cloud computing business, is hitting the gas pedal on its push into artificial intelligence in an unusual way. Andy Jassy, CEO of Amazon Web Services, announced AWS DeepRacer: A remote controlled car, retailing for $400 when it becomes available next year, that developers can program to drive itself. This car is now available to pre-order on Amazon at a special introductory price of $249.

This radio-controlled, four-wheel drive vehicle is 1/18th the size of an actual car. It’s trained using reinforcement learning, an AI technique that means the car will learn to drive better through trial and error. The car is rewarded for staying on the track, avoiding obstacles and more, helping the car to learn to drive more accurately over time…. Importantly, the same principles used to train DeepRacer could also be used to program real-life self-driving cars that could carry humans or cargo….” Read more Hmmmm…. Did Business Insider turn into The Onion??? Just because one might be able to learn in a vacuum by randomly correcting when one makes mistakes, this is a really naive, (I’m being nice) and not even close to “optimal”, way to acquire intelligence. Hopefully anyone that is seriously trying to program driverless cars is using more intelligence in their code. If, what is being developed is only for self-driving cars, they’d better make sure that the adult supervision always alert and able to intervene. For Driverless, this approach is sophomoric, at best. Alain