2013-10-17

October 17, 2013

October 18, 2013

Are You Ready When Your Google Car Freaks Out?

Junko Yoshida 10/16/2013 TOKYO — All eyes have been fixed on the autonomous car at the ITS World Congress here this week. For carmakers, regulators, and technology suppliers who have gathered here from all over the world, the big concern is not so much the bells and whistles of self-driving cars, but more about deployment strategies for them.

Debates here covered everything from human-machine interface, safety, and reliability to societal acceptance and a whole new legal landscape.

One issue, however, was artfully dodged by panelists on a Tuesday, Oct. 15, executive session called “Autonomous Vehicles – the Path to Implementation.”

Put simply, how does a self-driving car, when faced with a crisis beyond its program, hand control back over to the (human) driver?….At issue here is how well a self-driving car like Google Car can handle “exceptions” and how carmakers expect a driver to take over the control of the self-driving car in a critical situation.

Asked by EE Times, after the ITS session, about how self-driving cars are designed to handle exceptions, Ron Medford, Google’s director of safety for self-driving cars, deflected the question. He said, “Readiness of drivers to take over the control [of a self-driving car at the moment’s notice] is simply not understood yet.” Citing a number of studies now underway, Medford explained that, not only companies like Google, but also a lot of government officials and regulators are all eager for the results…” Read more Also be sure to look at the comments.

This same issue came up at the International Task Force On Vehicle-Highway Automation held the day before ITS. My comment was that Google must already have a pretty good idea about this issue. They’ve driven a half million miles. They’ve experienced a very large number of transitions between self-driving and human-driving. Their vehicles must have encountered “exceptions” when the driver had to take over “at the moment’s notice”. Granted the drivers are trained and exceptional (as we all think we are) so their behavior may not be a perfect representation of the distribution of driving behaviors that exists on the road today; however, they’ve gained an enormous amount of experience. Google should share those experiences with the human behavior research community so that it can build on the Google experience rather than having to redo it. Alain

Texas Instruments drives customer applications one lane closer to an autonomous driving experience and reduction in collisions

BADEN-BADEN, Germany, Oct. 16, 2013 “…Texas Instruments Incorporated (TI) (NASDAQ: TXN) today unveiled an automotive System-on-Chip (SoC) family, the TDA2x, incorporating an innovative Vision AccelerationPac with features to help customers create advanced driver assistance systems (ADAS) that help reduce the number of collisions on the road and enable autonomous driving experiences. Unlike any other processors in the industry, TI’s highly integrated TDA2x device family combines an optimal mix of high performance, vision analytics, video, graphics and general purpose processing cores in a low power envelope, enabling a broad range of ADAS applications including front camera, surround view and sensor fusion. Furthermore, TI’s unique Vision AccelerationPac complements the industry’s leading TMS320C66x digital signal processors (DSP) generation cores, and enables more ADAS algorithms to run simultaneously…” Read more Very interesting. Someone needs to jump on this and we need to get the SDK. Alain.

Empowering automotive vision with TI’s Vision AccelerationPac

“…TI’s Vision AccelerationPac is a programmable accelerator created specifically to enable the processing, power, latency and reliability needs found in computer vision applications in the automotive, machine vision, and robotics markets. The Vision AccelerationPac contains one or more Embedded Vision Engines (EVE) that deliver programmability, flexibility, low-latency processing as well as power efficiency and a small silicon die area for embedded vision systems. The result is an exceptional combination of performance and value. Each EVE delivers more than 8× improvement in compute performance for advanced vision analytics than existing ADAS systems at same power levels…” Read more ADAS Applications Processor TDA2x System-on-Chip Technical Brief

Consumers More Likely To Use Self-Driving Cars From Tech Cos. Over Traditional Automakers

“…The KPMG research yielded three important insights into when, why and how consumers might use self-driving vehicles.

• There’s a distinct self-driving value proposition. Get it right and consumers will clamor (and pay) for the technology.

• Get ready for the post-powertrain ecosystem. Acceleration time from 0-60 mph may not matter in the self-driving era. Consumers might well buy their self-driving cars from high-tech companies.

• The growth in self-driving mobility on demand services could mark the end of the two-car family.

“We realize significant hurdles and open questions remain, including safety, liability and even cyber-security concerns,” added Silberg. “In addition, technological innovation often moves faster than legal or regulatory systems. However, we believe the market opportunities for self-driving vehicles and technologies are enormous, and innovative companies will continue to drive the technology forward.”…” Read the original report: Self-Driving Cars: Are We Ready?

It is another very good report from KPMG to go along with their Self-Driving Cars: The Next Revolution Alain

On-Demand Transit System Summons Buses Via An App

“Kutsuplus, which means “call plus” in Finnish, is an on-demand minibus service run by Helsinki’s public transit authority. The service lets riders summon a bus with their smartphone and choose a start and end point. The riders can also choose to take a private trip or share a bus with others. A bus can accommodate up to nine passengers. If a rider opts for a shared ride, Kutsuplus uses an algorithm to determine the most direct route for all of the minibus passengers whose destinations may vary.

The service costs a little more than a bus ride, but less than the usual Helsinki taxi fare…” Read more

Hmmm, I wonder what the Helsinki taxi drivers think about the new competition??? Plus, why would anyone want to take the regular bus now that they can be chauffeured door to door. There must be a catch somewhere. Else every mode-split model would predict a shift of essentially all conventional bus riders to this door to door service. If not, there is no catch, then no need for autonomous taxis. Helsinki has ‘em for “little more than a bus ride” and is on the road to full employment. What a city; what a country! Alain

Toyota’s baby steps towards vehicle autonomy

“Toyota has announced it will begin installing semi-autonomous driving technologies in its cars from the middle of this decade.

As part of this week’s 20th Intelligent Transport Systems World Congress in Tokyo, the Japanese giant has revealed a Lexus LS research vehicle armed with its first strands of the technology, dubbed Automated Highway Driving Assist (AHDA) and Lane Trace Control, along with a pedestrian safety package labelled Pre-Collision System (PCS)… AHDA is an advanced adaptive cruise system, labeled “cooperative cruise” for the way it functions not by the normal radar but through wireless car-to-car communications. Cars equipped with the gear provide each other with constant real-time acceleration, coasting, braking and location updates, improving the precision of vehicle separation and eliminating the errors made by radar’s analogue spatial sensory systems…” Read more

Hmmm (or maybe don’t bother reading more), from the image that Toyota released above, its “baby steps” are not pretty. The car in the video above is much more to my taste. One wonders who might be impressed with all of the stuff hanging all over the Lexus. One also wonders about the Toyota continued commitment to wireless car-to-car communications as their approach to 2015. Given that the probability will be infinitesimal for a very long time (well beyond 2015) that there will be another car around with which the Toyota can communicate leaves one scratching his head to trying to figure out who would buy such a car in 2015. Oh well, I guess that Toyota hasn’t received the memo indicating that car to car communications might not be the right basis for Advanced Driver Assistance Systems. Alain

Crash-less cars? Toyota, Ford systems steer clear of pedestrians

Paul A. Eisenstein, The Detroit Bureau Oct. 13, 2013 “A Toyota test car demonstrates its Pre-collision System that uses automatic steering to avoid a mannequin during the Toyota Advanced Technologies media briefing in Tokyo on Oct. 10, 2013.

With pedestrian fatalities taking an unexpected rise in recent years, automakers are looking for ways not only to reduce the death toll but also quite literally steer clear of pedestrians in the first place.

Both Toyota and Ford just unveiled new systems designed to detect when someone might walk in front of a vehicle and try to prevent a collision, or at least lessen the impact.

The new Toyota Pre-collision System, or PCS, uses onboard sensors to scan the road ahead of the vehicle, issuing an alert if there’s a risk of a crash. But, the maker says, “If the likelihood of a collision increases, the system issues an audio and visual alarm to encourage the driver to take evasive action, and the increased pre-collision braking force and automatic braking functions are activated.”

If the driver does not react in time, the vehicle will automatically try to steer away from the pedestrian.

Ford demonstrated a similar concept last week in Europe, its Obstacle Avoidance technology relying on a mix of sensors, including a camera tucked behind the rearview mirror, to scan the road for vehicles and pedestrians. Like the Toyota PCS, it can steer out of the way if a collision seems imminent…” Read More. Interesting video of a test car but not yet available in the showroom. Also, it does this without vehicle to vehicle nor vehicle to pedestrian communications. Alain

Calendar of Upcoming Events:

http://www.sitaonline.org/conference.htm

October 20-23, 2013

Mirage Hotel

Las Vegas, Nevada

http://www.podcarcity.org/washington/

Washington DC Oct 23-25

http://www.cs.cmu.edu/%7ECVAD2013/

First International Workshop on Computer Vision for Autonomous Driving

Sydney, Australia December 2, 2013

The Ethics of Autonomous Cars

Patrick Lin Oct 8 2013, “If a small tree branch pokes out onto a highway and there’s no incoming traffic, we’d simply drift a little into the opposite lane and drive around it. But an automated car might come to a full stop, as it dutifully observes traffic laws that prohibit crossing a double-yellow line….” Read more Good article as are some of the comments that follow. I’ll add mine: Yes, ethics are really important here, but we also need to not be sophomoric. Laws, even traffic laws, are created and interpreted with human behavior in mind. They haven’t been written as deterministic absolutes. There are nuances. It is necessary that those who are writing the logic and code for these SmartDrivingCars (They are NOT autonomous, nor will they be in my foreseeable future.) understand that these laws are NOT absolutes and that they direct and constrain in a real-world context. Code that applies traffic rules rigidly and without regard for context will fail in the marketplace. If these cars are going to do some of the driving for us, their behaviors are going to have to meet our minimum expectations. Some of us actually rode with our teenagers when they began to drive. We pointed out mistakes, we pointed out that “Yes” you can cross the yellow line when there is a branch in the road and no car is coming. Code writers for the smart driving vehicles will build these kinds of cues into the system. Sometimes rule-breaking is the right choice on the road because our legal rules necessarily oversimplify to cover the generality of cases. The beauty of code is that nuances that cannot be captured in law can be accounted for in algorithms. The Smart Driving Car challenge is not an ethical challenge it is a computer code generating challenge. Alain

October 4, 2013

Car Fire a Test for High-Flying Tesla

BILL VLASIC Published: Oct. 3, 2013 DETROIT —”It’s an automaker’s worst nightmare: graphic video footage of one of its cars engulfed in flames after an accident. The driver told the police that he hit metal debris on the freeway before the Tesla Model S caught fire. In the case of Tesla Motors, the fire that destroyed a Model S electric car on Tuesday is a stunning reality check for a company that has garnered almost unanimous praise for its battery-powered vehicles. The fire, on a highway exit in Kent, Wash., poses a serious challenge for Tesla and, at the same time, prompts new questions about the safety of lithium-ion batteries in electric cars. Read more

Hmmm. Let this be a wake-up call to all in the SmartDrivingCar community that we need to be careful and be prepared to address accidents involving SmartDrivingCars. There will be accidents; nothing is perfect. We should/must continue to advance SmartDrivingTechnologies, take risks, be ready to keep everyone aware of the enormous benefits, and yes, nothing comes for free, not even Google; some prices will need to be paid.

What this article fails to bring out is that highway vehicle fires are “common”.

From Highway Vehicle Fires (2008-2010) FEMA Topical Fire Report Series Vol 13, Issue 11, Jan 2013:

“…an estimated 194,000 highway vehicle fires occurred in the United States each year (1.6 per hour!)
…an annual average of approximately 300 deaths (~1 per day), 1,250 injuries and $1.1 billion in property loss.
…highway vehicle fires accounted for 14 percent of fires responded to by fire departments across the nation.”

Special Issue

September 28, 2013

IIHS issues first crash avoidance ratings

IIHS News

Sept. 27, 2013 ARLINGTON, Va. — “A new test program by the Insurance Institute for Highway Safety (IIHS) rates the performance of front crash prevention systems to help consumers decide which features to consider and encourage automakers to speed adoption of the technology. The rating system is based on research by the Highway Loss Data Institute (HLDI) indicating that forward collision warning and automatic braking systems are helping drivers avoid front-to-rear crashes.

The Institute rates models with optional or standard front crash prevention systems as superior, advanced or basic depending on whether they offer autonomous braking, or autobrake, and, if so, how effective it is in tests at 12 and 25 mph. Vehicles rated superior have autobrake and can avoid a crash or substantially reduce speeds in both tests. For an advanced rating a vehicle must have autobrake and avoid a crash or reduce speeds by at least 5 mph in 1 of 2 tests… The Institute awards as many as five points in the autobrake tests, based on how much the systems slow the vehicle to avoid hitting the inflatable target or lessen the severity of the impact. In the case of an unavoidable collision, lowering the striking vehicle’s speed reduces the crash energy that vehicle structures and restraint systems have to manage. That reduces the amount of damage to both the striking and struck car and minimizes injuries to people traveling in them.

“We decided on 25 mph because development testing indicated that results at this speed were indicative of results at higher speeds — and because higher-speed tests would risk damaging the test vehicles,” Zuby says. “As such, we expect crash mitigation benefits at higher speeds as well.”

Hats off to Subaru for leading the pack in this first round of tests!

Be sure to look at the scoring table at the bottom of the IIHS news release. It is disheartening to learn that for the most part, these systems didn’t work! Only Subaru, Cadillac and Volvo didn’t crash in the 12 mph test and only Subaru in the 25 mph test. The purpose of these systems is crash avoidance! Each knew the crash was coming.

Why would manufacturers that took the effort to include automatic braking would wait until it is too late to avoid a collision or apply the brakes too lightly, allowing a crash to occur. Even a slight crash causes a high “cost” (least of which requires you to pull over, talk to the person that you just ran into); whereas no crash incurs zero “cost” (except an elevated heart beat). Alain

September 27, 2013

House to hold hearing on driverless cars

By Keith Laing - 09/25/13 “The House Transportation and Infrastructure will hold a hearing next month about “the future role of autonomous vehicles in U.S. transportation,” officials with the panel announced on Wednesday. Transportation Committee Chairman Rep. Bill Shuster (R-Pa.) proclaimed driverless cars were “the future of transportation” after riding in one in his home state earlier this month. “This technology has significant potential to make transportation safer and more efficient,” Shuster said in a statement after his ride. “We have to figure out how to embrace technology, in the way we build our infrastructure, comply with existing and future laws, and ensure the safety of the public.” Read more

Hmm…but there are no details as to when and whom they’ve requested to appear. It does seem to have an appropriate positive thrust of “… how to embrace technology…” Also, driverless cars is an unfortunate label for all of this. “All” of the safety and comfort benefits and most of the convenience benefits of SmartDrivingCar technology are likely to be captured without ever getting to driverless (Level 4). In fact, driverless (Level 4) will likely exacerbate some of the safety benefits gained by Levels 1 through 3 in return for very modest gains in personal convenience (It’ll send my car back home for my spouse to use, or fetch my car from some remote parking lot. Not a big deal!?).

The benefit of driverless is captured not by the individual car buyer/owner, but by the emerging “mass transit” company that has purchased a fleet of these vehicles for the purpose of making them available to the traveling public on a “per ride” basis. This corporate entity will manage its fleet to best serve its potential customers without incurring a labor cost with each vehicle. Conventional mass transit does something close to this with large vehicle carrying lots of people. This is the only way they can keep the labor cost per individual customer at an acceptable level. These new “mass transit” firms, call them autonomous Shared Taxis (aTaxis), will be able to afford to offer mobility services using much smaller vehicles, each carrying many fewer people to a degree that they could offer inexpensive Point to Point on-demand mobility services to all. This would make “mass transit” competitive and possibly the mode of choice for many more, if not most trips. If so, investors would find it attractive to acquire sizable fleets of driverless vehicles and offer “mass transit” mobility that could replace most current personal auto mobility. The enhanced ride-sharing that would naturally emerge from such offerings could substantially reduce congestion, energy and pollution while not substantially degrading safety. Creating a welcoming environment for such investments in driverless vehicles that have the potential to yield very valuable societal benefits should be a major focus of these hearings.

To achieve the “safety” objectives, the hearing need only to “embrace” Levels 1-3 technologies. Here “embracing” may only need to make sure that “driverless hurdles” (Level 4) are not placed as roadblocks to Levels 1-3 market availability. Alain

September 20, 2013

At Frankfurt Auto Show, the Driver Began to Take a Back Seat

By JACK EWING Sept. 15, 2013 FRANKFURT —” A wide grin beneath his bushy mustache, Dieter Zetsche, the chief executive of Daimler, did as car executives often do at auto shows, cruising onto the stage in the company’s newest model. But at the Frankfurt motor show last week, Mr. Zetsche added a surprise: he sprang from the back of a Mercedes S-Class that had no one in the driver’s seat…” Read more This is how Daimler chose to spend a substantial amount of money to introduce its automotive products at the 2013 Frankfurt Auto Show on Sept. 9, 2013. They must believe that consumers are ready to spend money on Smart Driving Cars. Alain

Video: MB Self-Driving Manheim 2 Pforzheim 2:08 long

S 500 MB Intelligent Drive (Self-Driving) TV footage:

September 13, 2013

Race is on as automakers embrace self-driving cars

David Shepardson September 12, 2013 Frankfurt, Germany — “ The race to build an autonomous car by 2020 is on. Several global automakers at the international auto show here vowed that they are in the lead — and trying to one-up each other…. Mercedes-Benz — which touts its safety advances — is ramping up its efforts as it adds additional crash mitigation technology to current vehicles.

“Those who want to drive themselves are free to do so, and that won’t change in the future, either,” Weber said…” Read more

Well worth reading Alain

2013

A New Product Category: the PCwD !

Personal Collision-warning Device (PCwD) Available at under $850

The market for Collision Warning Systems may well be developing in a manner similar to that of turn-by-turn navigation where consumer-grade after-market systems are cheaper and better than “factory installed” versions sold as options with new cars. The popularity and adoption of navigation systems was fueled by their availability to all car users as inexpensive stand-alone Personal Navigation Device (PND) that could be used in any car or as less expensive, even “free”, application programs that leverage the capabilities of one’s existing device, be it a notebook, tablet, iPad, or smartPhone. Mobileye is out with what I’d like to call a “Personal Collision-warning Device” (PCwD). It is a self-contained unit that uses embedded stereo cameras to monitor the road ahead and warn the driver of forward collision, pedestrian collision, lane departure and read speed limit signs. At a price of just under $850 it is substantially cheaper than comparable systems available as options on new cars. More importantly, the product is available to everyone, not just buyers of a few models of new cars. Given the universal desire to “text” and the allure of other distractions while driving, everyone needs the help of a PCwD to be constantly vigilant of potential dangers ahead.

While this is the first such device, I see Collision-warning Systems following the market adoption path blazed by Navigation Systems. The main difference here is that stereo cameras have replaced the GPS receiver. The user-interface, computing and memory are each essentially the same. Other producers with their own image processing (routing software), object recognition (digital maps) and user interfaces will emerge and propel the market adoption of PCwDs. At some point, communications with some evolution of the “OBD II” port could be established with the a car’s electronic throttle, brake and steering systems to create a PCaD that provides active Collision-avoidance (Ca) and lane-keeping rather than simply collision-warning.

The final step would be to merge the PCaD with the turn-by-turn navigation system to create the ultimate new product category, the Personal Driverless-car Device (PDcD) that could behave as a Valet or as a Chauffeur and allow you to catch a few zzzz. Whew! Alain

See following videos for more information:

2013_Commercial_Teen About Mobileye Links to Videos of individual capabilities available @ https://us.mobileye.com/products/mobileye-560/

alaink@princeton.edu