Monday, February 3, 2014

Did NHTSA say anything new on a connected vehicle mandate?

The National Highway and Traffic Safety Administration (NHTSA) has been studying the benefits of connected vehicles for decades. This culminated in the Connected Vehicle Testbed Model Deployment in Ann Arbor, where a fleet of connected vehicles hit city streets in 2012. NHTSA's end-goal has long been to standardize and mandate a particular type of connected vehicle technology, known as direct short-range communication (DSRC). The Ann Arbor deployment was designed to acquire data to support such a regulatory decision.

On February 3, 2013, NHTSA via formal press an intent to "move forward with vehicle-to-vehicle communication technology for light vehicles." In the center of boilerplate press language about the benefits of V2V, this seems to be the meat of the announcement:
NHTSA is currently finalizing its analysis of the data gathered as part of its year-long pilot program and will publish a research report on V2V communication technology for public comment in the coming weeks. The report will include analysis of the Department's research findings in several key areas including technical feasibility, privacy and security, and preliminary estimates on costs and safety benefits. NHTSA will then begin working on a regulatory proposal that would require V2V devices in new vehicles in a future year, consistent with applicable legal requirements, Executive Orders, and guidance.
Yeah, OK then... So NHTSA is studying V2V technology. We knew that already.

This press release suggests that a the formal regulatory process has not yet even been initiated. NHTSA has not even yet begun work on a regulatory proposal? They didn't mention posting anything in the Federal Register. It seems like nothing has changed since yesterday. Am I reading this wrong? What, exactly, did we learn from this press release?

NHTSA announcement is here: http://www.nhtsa.gov/About+NHTSA/Press+Releases/2014/USDOT+to+Move+Forward+with+Vehicle-to-Vehicle+Communication+Technology+for+Light+Vehicles


Wednesday, January 22, 2014

Barriers to Successful Implementation of a National DSRC Connected Vehicle Network

The following is an extended abstract that I have submitted as a draft technical paper for the ITS World Congress 2014.

ABSTRACT

The primary goals of the U.S. DOT connected vehicle research are to improve surface transportation safety, mobility, and environmental performance in the United States. As of this draft technical submission submitted January 22, 2014, the ITS community awaits a National Highway Traffic Safety Administration (NHTSA) decision regarding direct short-range communication (DSRC) connected vehicles in the United States. Many stakeholders anticipate that NHTSA will announce an intent to mandate that DSRC equipment be installed in all new vehicles at a future date. However, a critical assessment of this strategy suggests that there potential complications that must be considered. This paper provides an overview of procedural, political, and technical barriers that impose significant difficulty in pursuing a DSRC connected vehicle mandate.

U.S. DOT connected vehicle research can be traced back to the early 1990s, when the Department commissioned the development of a strategic plan for Intelligent Vehicle Highway Systems (IVHS) in the United States. The resulting report was prepared by IVHS America—an official congressional advisory committee—and was published in May of 1992. The strategic plan outlined a 20-year program by which a partnership of federal, state, and private agencies would work together to design and deploy a national IVHS program for increased safety, improved mobility, and environmental quality. Such goals proceeded through the U.S. DOT's Automated Highway System (AHS) demonstration, the subsequent Vehicle Infrastructure Integration (VII) program, and lives on today in the strategic research program managed by the ITS Joint Program Office within the Research and Innovative Technology Administration (RITA) of the U.S. DOT. These federal initiatives have successfully supported the standardization, technological advancement, validation, and deployment of numerous ITS systems across the U.S.

Most existing ITS deployments in the U.S. today are managed by local or state transit agencies, and often operate independent of adjacent or overlapping ITS deployments. Many ITS industry stakeholders believe that the greatest potential for returns to investment on ITS installations will require that ITS systems are integrated into a national network of connected vehicles and infrastructural components. The U.S. DOT has adopted such a view in its Statement of Principles for a Connected Vehicle Environment which reads, "System implementation must be national in scale and extensible across North America." The system is envisioned as a centrally managed national network provided as a public service using open nonproprietary communication and performance standards. The U.S. DOT has found that DSRC is the only known viable technology for safety critical applications provided within such a system. While the Department's approach to pursuing the goals of the connected vehicle program is continually evolving, many ITS and safety advocates have called on NHTSA to pursue a mandate requiring DSRC standard equipment be installed in all new vehicles sold in the United States.

A critical analysis of a DSRC mandate scenario indicates that such an approach may be fraught with difficulty. The most immediate barrier to a DSRC mandate concerns limitations on the U.S. DOT and NHTSA as regulatory agencies. Pursuant to federal requirements, regulatory agencies are subject to a critical review process to ensure that regulations are designed to achieve benefits that justify any costs imposed. Advocates of the U.S. DOT connected vehicle program argue that the cost of a mandate is likely to be marginal when compared to the potential savings in fatalities, injuries, property damage, lost economic activity, etc. NHTSA estimates that motor vehicle crashes in the United States have a negative economic impact of about $230 billion each year (including thousands of serious injuries and lives lost) and that up to 80% of these crashes could be prevented with connected vehicle technology. However, the results of a formal cost-benefit analysis are far from certain. This is partially a result of established accounting principles that emphasize near-term costs and benefits; a national DSRC network is likely to require significant upfront cost with the bulk of benefits accruing years or decades later. Long-term impacts of a DSRC mandate would be subject to a discount rate that will significantly depreciate the value of projected benefits. Further, many claims about the safety gains of a DSRC mandate may not stand up to critical review. Even under favorable assumptions, the installation of DSRC equipment alone is unlikely to have much safety impact unless coupled with automated safety systems. Many automated safety systems rely on sensor technology and can operate independent of DSRC support. Such systems already appear to be contributing to a downward trend of vehicle crashes in the U.S. The safety gains of DSRC would likely have to be compared to baseline assumptions of safety increases provided by alternative technologies, further reducing the estimated marginal benefits provided by a DSRC mandate. In addition to the likelihood that the benefits of a DSRC mandate would be less than proponents expect, it is difficult to envision NHTSA being capable of estimating the costs of such a mandate with much precision—as some very basic elements network design and management remain unknown.

If NHTSA is able to show that the benefits of a DSRC mandate outweigh the costs, the successful adoption of such regulation may require extensive political capital. The current political climate on the United States is such that government involvement that appears reasonable to some may appear onerous—even nefarious—to others. Any proposed regulations regarding a DSRC mandate in light vehicles will likely attract a great deal of public attention. Special interest groups and media sensationalists will compete to control the narrative. Potential adversaries of a DSRC mandate include civil liberties organizations and advocates of limited government. One possible talking point of DSRC adversaries is that the Government wants to put a tracking device in everybody's car. And to be fair, this is partially true. Placing aside the technical difficulties of ensuring that DSRC equipment is secure and motorists are not tracked, the rhetorical difficulty of convincing the public that privacy will be maintained will be significant. The U.S. DOT and NHTSA would be well advised to consider such broader implications when considering regulatory mandates.

Beyond barriers relating to proposing and adopting a DSRC mandate, the potential difficulties for implementing a mandate are considerable. Current conceptions of the potential DSRC connected vehicle network heavily emphasize security and privacy, which adds complexity to the system. Specific security standards remain in development, but are expected to include a public key infrastructure (PKI)—essentially encoding each DSRC transmission with an authentication ID code that identifies the message as coming from a legitimate device. In order to protect the privacy of drivers, it is expected that each vehicle will update its ID code periodically so that the code cannot be used to track the location of the vehicle for a significant time or distance. The maintenance of the PKI system, as well as the general operation of the network, will require complex engineering and intensive IT support. The cost of deploying and running the system is very difficult to estimate at this stage in development but is likely to be substantial. The deployment and operation of such a network may amount to an IT project at a magnitude and complexity that is without precedent. Further, the scope NHTSA's authority does not likely include the operation of such a system, implying that novel public-private management structures may have to be designed to deploy and operate the network. This management dynamic may be very complex and plans for a supporting institutional structure remain very preliminary.

This paper provides a critical analysis of the prevailing direction of U.S. DOT connected vehicle efforts, focusing on a potential DSRC mandate. It is found that NHTSA intent to propose a mandate would involve a variety of procedural, political, and technical barriers. The culmination of the various potential barriers to a DSRC mandate for connected vehicles suggests that alternative regulatory approaches should be evaluated in earnest


Friday, December 20, 2013

Google's Purchase of Boston Dynamics Could Fast-Track Automated Driving

Google has been working on artificial intelligence (AI) and machine-learning algorithms for many years. By acquiring Boston Dynamics, Google has not only purchased terrifying proto-Terminator robots, but they have gained a plethora of AI that has been specifically developed for robotics.

The broadened catalog of AI algorithms will likely benefit Boston Dynamics legacy programs, but has huge implications for one of Google's most hyped projects: the Google Autonomous Car. The Google Car has already logged millions of miles in traffic situations and appears to be quite reliable on highway driving. This is impressive, but not unique. Many traditional automakers have already commercialized automated driving systems capable of limited self-driving (or partial-automation) in a highway environment. Automated an entire trip is more challenging by an order of magnitude. It's relatively easy to program a car not to crash, but it's not at all clear how to program it to properly navigate congested surface streets around unpredictable traffic, pedestrians, obstructions, etc. What Google and others have realized, is that it is much easier not to program these vehicles, but to train them.

Google and others have become remarkably good at the mechanical aspect of automated driving. Driving is only partially mechanical; it is largely social. We don't realize how complex of a task driving is because we have been evolving algorithms for millions of years that allow us to navigate both our physical and social environments. You can easily program an automated vehicle to obey the rules of the road and not hit things. But unless the vehicle can interact in our social world in a relatively human way, it will become confused and freeze up while trying to navigate in our world.

The conventional wisdom is that Google is using its fleet of automated vehicles to test its automated driving system. This is certainly true, to an extent, but it's not the most important aspect of what Google engineers are doing. Whenever the Google car encounters a unique or unusual situation that the software is not ready for, the test driver must temporarily take over the dynamic driving task. This is important for safety reasons to test in public, of course. But this is important for program development, because the car is watching and learning from the human driver. The aggregated experience of the fleet of Google cars is used to update and refine the controlling algorithms. Software programmers are surely involved in this process, but they don't have to program each unique traffic maneuver from scratch. They have they data regarding what the cars sensors saw in a unique situation, and can see how and why the algorithm became confused, causing the driver to take over. They can use this information, in addition to what the driver did, to increase the capability of the software. Boston Dynamics very likely has a lot of technology that will improve the ability of the Google automated driving system to learn from experience.

A year or so ago, I believed Google was basically playing with toys here. Or maybe the technology would be ready decades from now. Every time I've revised my view on this, it has involved increasing my respect for the Google Car program and subtracting from the time I would predict the technology will be ready. Yeah, OK then... My current prediction: I believe Google will have a consumer-ready driverless (NHTSA level 4 automation) product or service before 2018.



.

Tuesday, December 17, 2013

Michigan Considering Ban on Automated Vehicles

As I previously reported, Michigan's state legislature has been working to make Michigan the 4th state to pass a law concerning automated vehicles. The bill has gone through some revision and now passed through congress as SB 0169 of 2013 and is ready to be signed by the governor. While the intent is to highlight Michigan's potential as a high tech state, a significant effect of this law would be to make Michigan the first state to explicitly ban automated vehicles for non-testing purposes. This may sound reasonable for those who think we're still several years away from road-worthy driverless cars. However, the way the bill defines automated vehicles, some of today's cars may fall under the law.

Here is the S.B 169 language; emphasis is mine.
SEC 2B (1) "Automated motor vehicle" means a motor vehicle on which automated technology has been installed, either by a manufacturer of automated technology or an upfitter that enables the motor vehicle to be operated without any control or monitoring by a human operator. Automated motor vehicle does not include a motor vehicle enabled with 1 or more active safety systems or operator assistance systems, including, but not limited to, a system to provide electronic blind spot assistance, crash avoidance, emergency braking, parking assistance, adaptive cruise control, lane-keeping assistance, lane departure warning, or traffic jam and queuing assistance, unless 1 or more of these technologies alone or in combination with other systems enable the vehicle on which the technology is installed to operate without any control or monitoring by an operator.

(2) "Automated technology" means technology installed on a motor vehicle that has the capability to assist, make decisions for, or replace an operator.

(3) "Automatic mode" means the mode of operating an automated motor vehicle when automated technology is engaged to enable the motor vehicle to operate without any control or monitoring by an operator.
...

Sec. 663. Except [for testing purposes], a person shall not operate an automated motor vehicle upon a highway or street in automatic mode

So SB 169 explicitly bans operation of a vehicle in automatic mode. In other words, outside of testing scenarios, one cannot use features of a vehicle that can "operate without any control or monitoring by a human operator."

The language of the bill is sloppy enough that this could be interpreted such that no driverless car could ever be banned. Alternately, a few 2014 model-year vehicles could be illegal to operate. This is because the legislature tied crucial definitions to words like control and monitoring, but did not define these terms.

On one extreme, this bill may not ban any automated vehicle ever. Even a completely driverless car with no human inside would ever be said to be out of control unless it is malfunctioning. It is under the control of the software and software designer, and most likely could be remotely controlled, operated, and monitored in a variety of ways that would satisfy such a legal requirement.

On the other hand, we may adopt a very strict definition of control would require a driver to have hands on the wheel and/or feet on the brakes, or at least ready to do so in a moment's notice. Legacy vehicle codes use the language "actual physical control" to describe what it means to operate. By dropping that language, it seems SB 169 recognizes that control may not require physical contact between a human person and a car, but seems to suggest that legislators assumed a type of control where a human operator could instantaneously assume actual physical control at any time.

It is also critical that in SB 169 automatic mode occurs if the human driver gives up control OR monitoring, but not necessarily control AND monitoring. In other words, if you watch your car control itself, you are in automatic mode. This is the biggest problem, because there are at least a few 2014 vehicles that would suddenly be illegal to drive (or monitor the driving of) if SB 169 becomes law.

Here are three vehicles that could be banned from Michigan under this law:

1. The Acura RLX


For about $12,000 over the base model, the 2014 Acura RLX can be purchased with the Advance Package. This slew of driver assistance, convenience, and comfort features includes advanced adaptive cruise control with low-speed following, collision mitigating breaking, and lane keep assistance. The combination of these features in highway driving would allow highway driving with minimal if any monitoring by a human operator.

2. The Mercedes Benz S-Class
Front view 2014 Mercedes-Benz S-Class

Mercedes offers the 2014 S Class with an optional $2,800 Driver Assistance Package that allows the car to be placed in [what SB 169 would call] automatic mode in stop & go highway traffic. The automatic mode can also be used at high speeds, though only for 15 seconds at a time.

3. The Infiniti Q50


Infiniti's 2014 Q 50 can be purchased with an optional $3,200 Technology Package that includes full speed intelligent cruise control, active lane control, and forward emergency braking. Infiniti does not label this automated driving, but it is hard not to interpret this combination of features as providing an automatic mode, or what NHTSA would define as Level 3 automation, limited self-driving. A New York Times reviewer observed, "I found myself driving the Infiniti on surprisingly long highway stretches without touching the accelerator, brake pedal or steering wheel."

From would I can tell, it wouldn't be illegal to OWN these vehicles in Michigan, but it would be illegal to use these optional features. Maybe Ford or GM should point out that they are planning to bring similar technologies to market within the next few years.



Thursday, October 24, 2013

A Review of KPMG's White Paper "Self Driving Cars: Are We Ready?"


This is an informal review and critique of the KPMG white paper titled, Self-Driving Cars: Are We Ready? The KPMG paper has gathered some attention due to its finding that consumers would be more willing to buy a driverless car from Google than from traditional automakers. This caused a bit of a stir in Detroit, considering Google has never made a car.

The primary methods used to gather evidence for concluding statements made in this paper includes consumer focus groups and KPMG's Mass Opinion Business Intelligence (MOBI) data analysis methods. Cursory historical and background research was also included.

Focus Groups

The KPMG paper reports conducting three focus groups in June 2013 in Chicago, Los Angeles, and New Jersey. The given demographics of focus group participants is given as follows (pp. 16):
  • Age: 21 or Older: 100%
  • Family Income: $50,000/year or Greater: 100%
  • Vehicle Ownership: Own and have purchased within previous seven years: 100%
The demography of focus groups is not broken down any further. In addition, the size of the focus groups is not given individually or in sum. It is noted that focus group insights are "not statistically valid" and are important for "qualitative, directional insights" (pp. 14).

Focus group participants reported passion for driving on a 1-10 scale, along with usual driving characteristics (i.e., mileage per week, highway or suburban, etc.). They were then shown a YouTube video featuring a blind man piloting a Google prototype automated vehicle, and asked to rate how likely they would be to use such a car on a daily basis. They were twice subsequently asked to re-rank their desire to use an automated vehicle based on information disseminated over the course of the discussion. It was generally found that participants were more willing to use a driverless car at the end of the discussion.

Five individuals were selected and profiled to represent a cross-section of participants and lend specific context to some of the findings. If this was done honestly (real people actually said quotes attributed to them), I think this is legitimate and useful for a market study. I think that many of the items that CAR affiliates may find controversial are attributed to the participants.

I think that when focus groups are done properly, they can be useful in generating research directions and exploiting a base of experience, knowledge, and perception that may otherwise be untapped. I have some concerns with this particular use of focus groups and would be hesitant to make any statements about market viability of automated vehicles based on this.

While it is stated that results of surveys conducted during focus groups are "not statistically significant," the results from surveys are presented as quantitative data. As quantitative data is presented, I think that there should be a discussion of sample size. I think the preferred approach would have been to use the focus groups to generate research questions, but present it as such, not as data.

Five participants were profiled and provide a narrative base to lend support to the market readiness for these automated vehicles. I believe this was done with the theory that people respond to stories more than statistics. This theory of communication may be true, but I don't think there's any real information in these profiles. They are of no use to me as a systems analyst. However, if I were a market analyst, they may be informative. The greatest concern to me regarding the use of these focus groups is that participants seem to have received unrealistic descriptions of hypothetical products that would be coming into the marketplace. The Google video they were shown was a selectively-edited promo video itself. They were also told that these cars would reduce commute times significantly and may be given special automated-only lanes. I don't see any compelling evidence of even reasonably near-term possibilities on automated-only lanes or reduced commute times. So, what these focus groups mainly found was that when presented with an idealized hypothetical product with unrealistic benefits and minimal costs, consumers were positive. That said, the stated goal of this paper was to determine market readiness for a fully-automated car, which does not exist; so I'm not sure how else one would go about this.

MOBI

I'm not sure how to assess the MOBI contribution. I'm generally supportive of big data analytics, but they don't give any details about this method. I think there's probably something valuable in it, but I don't have enough information about the method to know what the use of the MOBI data is. That said, MOBI data is not used to make any claims I would be inclined to reject.
Claim: Online discussion of driverless cars is at an all-time high (pp. 6).
I'd agree.

Claim: Online discussions regarding automated vehicles focus on different features than standard automobiles (pp.17).

The MOBI data suggested that when compared to what people are concerned about with normal cars, concerns about automated vehicles tend to emphasize the following:
· Handling
· Safety
· Innovation
· Trust
· Quality
· Insurance

The following factors are deemphasized for automated vehicles:
· Engine
· Transmission
· Style
· Interior

I think this conclusion makes sense, particularly when the definition of handling is expanded to include, basically, driving "skill."

Claim: Automated vehicle brand-association favors Google (pp. 28)


MOBI was used to determine that automated vehicles were most closely associated with Google as compared to other brands. Google is followed by Nissan, Mercedes, then everyone else. I have no cause to doubt the accuracy or legitimacy of this finding. Google had a marketing blitz. Nissan and Mercedes have also been relatively open about their research activities and have also had marketing efforts. It's only been very recently that domestic manufacturers have been forthcoming about automated vehicle technology development.

I think there may have been a mistake in tone here, or maybe just formatting. The sub-heading above the relevant data chart is "Which Brands Will Matter?" I think it's premature to assume that Google automated vehicles will "matter" more than others because they currently get the most Google-hits. That said, I do not discount Google's ability to enter and compete in the vehicle market (as many auto-industry insiders do).

Traditional Research

Research based on third-party sources did not play significantly into the conclusions of the paper, but was included. Pages 7 and 8 outline technological and regulatory updates in the last year. I was a silent contribute to this section via my work at the Center for Automotive Research in Ann Arbor, MI.

A "historical perspective" is also offered (pp. 12-13). This briefly discusses the uncertainties associated with innovation and features the transition from horses to automobiles. I think this comparison is overemphasized, but not inappropriate.

On page 30, there is a research-based discussion predicting a future vehicle market. It is suggested that the ownership model may be transition to a mobility-as-a-service model. This is a popular discussion now and I think there is some truth to it. I don't think the market will change quickly or drastically. But all this paper suggests is that the ratio of mobility sales may change; trending towards service and away from ownership. It's premature to make this prediction, but I think it is properly hedged and placed in context. I do not disagree with anything regarding this discussion. I also believe there will be a growing market for mobility-as-a-service.

Review of Conclusions

The paper itself claims that the "crux" of the research is these three insights, which will be discussed in sequence: 
  • Focus on Improving Consumers' Quality of Life 
  • Prepare for Radically Different Automotive Ecosystem 
  • Expect New Threats and Opportunities from Mobility and Demand Services

Focus on Improving Consumers' Quality of Life

It is claimed that consumers' receptivity to fully-automated vehicles increases significantly when presented with a product that features:
  1. shorter commute times
  2. reduced traffic-related variability
  3. ability to turn automated feature on/off
It is concluded that the market may be dominated by a company that can deliver on these features with an "esthetically and emotionally pleasing" experience. I agree, but I do not think this is a reasonable expectation. It would have been more useful to explore market readiness for near-term solutions. For example, several auto companies will be offering low-speed full automation in MY2014 packages for about $3,000. However, state laws are not ready; can drivers expect to be able to use personal devices, read books, etc. when the system is engaged? Would consumers risk the ticket?

Prepare for a Radically Different Automotive Ecosystem

It is claimed that future vehicles might be manufactured with minimal attention to performance (speed, handling, etc.). It is even suggested that tech companies with offer products or services with their brand.

I think this conclusion is premature, but not off base. I would hesitate to publish such predictions because I don't think the evidence is there, but I don't totally disagree. I think that "ecosystem" change will not be radical; it will be slow and quiet. However, I think it probable that tech manufacturers will begin offering branded mobility services based on fleet-owned automated vehicles. Rumors have begun leaking out that Google is gathering a supplier base. A car designed ground-up to be driverless has advantages, as does offering the driverless car as a service rather than product. I would not be surprised to see a Google-branded automated vehicle service launched in the Bay area within about 5-10 years. There is a market just getting their employees to work.

Expect New Threats and Opportunities from Mobility on Demand Services

It is claimed that mobility-as-a-service could cut into the vehicle-ownership market share. This is another claim that I would be hesitant to publish, but agree with. There is some data that suggest a movement away from vehicle ownership, but not enough to conclude a trend. Further, the mobility-as-a-service market is still small, and there is no automated mobility-as-a-service market as of yet. But if I had to guess one way or another, I would guess that there will be a slow but consistent trend towards mobility-as-a-service and away from vehicle ownership.

Review of Potentially Controversial Statements

The KPMG paper has not been widely noticed. But auto industry insiders from traditional OEMS who have noticed it were irritated at some of the findings and the general tone. Related to this reaction, I have scanned the paper for statements that may be interpreted as controversial, and briefly responded to each. I did not feature statements that simply make a statement that is sourced, (e.g., "Tesla Motors is talking about potentially adding self-driving technology to its vehicles.") since I just see this as relaying relevant information. I've also omitted statements regarding the conclusions ("important insights"), as these are covered above.
"Consumers might-well buy their next car from a high-tech company, such as Apple, Microsoft, Samsung, Google or Intel, among others" (pp. 4). 
This statement seems to draw support from the focus group and MOBI data. Since it is presented as a market possibility and not a trend or statement regarding the quality of traditional automakers, I don't believe it should be taken as controversial. I don't envision tech companies selling cars to consumers as this statement suggests. However, I do think that mobility-as-a-service under the brand of tech companies is possible. 
"The advent of safe, reliable fully self-driving vehicles will require the convergence of ... sensor-based solutions and connected-vehicle communications" (pp. 6). 
There is increasing evidence that automated vehicles can be successfully deployed without V2V. 
"Self-driving cars are not explicitly prohibited in any existing state laws, and some commenters have reasoned that they are, therefore, legal in all states" (pp. 8). 
This is my language based on my research and I stand by it until litigation proves otherwise.
"What about OEMs who don't move fast enough? Will they become extinct" (pp. 10)? 
I think this was irresponsible prose, but it is just posing a question. I also think OEMs should be aware of staying on top of trends to avoid severe market erosion. 
"In spite of having little or no experience in manufacturing cars, technology brands are trusted as much as premium auto brands" (pp. 15). 
This statement was extrapolated from focus group discussions regarding a hypothetical product, so was possibly irresponsible. I also think that KPMG intended to find this finding going in for whatever reason. However, I don't think they had to completely manufacture the sentiment. Whether deserved or not, innovation and quality is associated with high-tech brands far above auto industry companies. The statement was irresponsible but not necessarily incorrect. If this statement were to be substantiated, I think it's a salient warning to the auto industry that, yes, there is a possibility that tech-firms will be increasingly moving into the auto/mobility business. 
"The discussion revealed a noticeable bias against American Cars within our focus groups" (pp. 29).
If this conclusion was reached honestly, and not prompted, I think this is a legitimate statement and does not surprise me. The actual quality of American cars is never derided, only focus group perceptions. It may have been irresponsible to publish statements based on such a small sample size, but it might be appropriate for such a market-type study. Anyways, I do think that a proper survey would support this statement. I think American brands to have a brand-identity issue especially in some areas (think coasts) of the U.S. They do say that, "given that 2/3 of our participants came from the east and west coasts and large urban markets, our data is not necessarily reflective of broader national trends" (pp. 29).
[Focus Group Participant Quotations:] 
In the interest of brevity, I can sum up that the five participants who were profiled and quoted at length were generally derogatory towards domestic OEMs, with some exceptions. Brands that were spoken of favorably include Google, Apple, Mercedes, BMW, and Jaguar. No specific American brands were mentioned even in passing. I have no reason to believe that reporting of focus group discussions were biased to such statements. In fact, I suspect that the couple comments quoted in defense of the American auto industry were consciously highlighted. The question of whether or not it was appropriate to publish the findings in this format may be different. I am concerned about the small sample size and questionable methodology (e.g., beginning discussion with the Google promo video).

Conclusion

The difficulty in assessing this paper is that its stated intent is as market research for a market that does not exist. I do not know to what standards it should be held to. Several conclusions were made without much supporting evidence. As a technology/policy-based analyst, I would have been extremely hesitant to make many of the statements and conclusions reported by KPMG. That said, no statements strike me as drastically irrational or incorrect—just unsupported. The paper was highly skewed to reflect positively on technology companies over traditional auto manufacturers. Also, premium and non-U.S.-based OEMS are shown in a more positive light than domestic OEMs. Assuming that the report language accurately reflected good-faith efforts to capture market sentiment from focus groups and data analysis, I don't find the language of the paper to be inappropriate for a market study. I would guess that many of the "controversial" statements in this paper would be very familiar to internal market-research people at domestic OEMs.