|
Presentation by the Automobile Protection Association - May 9, 2017.
George Iny, Executive Director
Senate Standing Committee on Transport and Communications Special Study on the Regulatory and Technical Issues Related to the Deployment of Connected and Automated Vehicles
Honourable senators,
Thank you for the opportunity to present the Automobile Protection Association’s perspective on issues related to automated vehicles. To a certain extent, many of the vehicles we drive today already offer a limited form of automated driving, be it in transient emergency situations via features like electronic stability control, antilock brakes, or convenience features like automatic cruise control and self parking. Some of these features have existed for two decades; electronic stability control whose introduction dates back to the 1990s has attained an enviable record for collision reduction. These devices were developed by the automakers as add-ons to existing vehicle technologies and designed to work with them. The legal regime governing them was also unchanged; for example, the driver is still considered the primary party responsible for guidance of a self-parking car in the event of a collision or other incident, and the feature has not led to changes in insurance policies or premiums.
Planning for the advent of much more autonomous vehicles involves taking stock of the legal and regulatory framework we currently have, which has been tested and provides some predictability, to determine how it can be adapted, going forward.
We should also look backward to other precedents. The elevator, reported to be the safest human transportation device, underwent large-scale automation in the years after World War II. The elevator uses a basic algorithm that will determine what floors it will stop at along the way with no human operator. Today it seems totally normal when you enter an elevator, to press a button that will take you to your destination; one day we may be able to get into a vehicle with much the same expectation.
The situation today
The human driver has a poor reputation. Auto collisions are a leading cause of injury and death among the young, and human factors contribute to over 95% of collisions, usually as the primary cause. Lately, improvements in occupant protection have been holding the line on injury and fatalities, rather than reducing them as had been expected. The cause is generally attributed to declining attention behind the wheel from motorists, and more congested urban roadways that now feature a greater presence of non-auto users (pedestrians and cyclists). Intelligent vehicles that see when the human is not paying attention carry the promise of a major leap forward in collision reduction.
The current legal framework for safety defects
Federal law requires the auto manufacturer to “give notice of any defect in the design, manufacture or functioning of the vehicle or equipment that affects or is likely to affect the safety of any person (s.10 of the Motor Vehicle Safety Act).”
Federal law requires that companies… provide the Minister with the means to retrieve and analyse information created or recorded by the vehicle or equipment. (s. 8 of the Motor Vehicle Safety Act)
Both provincial and federal laws require that the person with “care or control of a motor vehicle […], whether it is in motion or not” is the first party responsible for its safe operation.
Manufacturer as owner of private vehicles
Millions of automaker-owned vehicles are already in private hands as a consequence of auto leasing. The lessee is required by the automaker to carry sufficient insurance to more or less ensure that the automaker will not suffer a loss arising from the vast majority of vehicle collisions.
Privacy
Most new vehicle owner’s manuals now explicitly recognise that onboard vehicle data collected in the minutes or seconds before a crash belong to the vehicle owner. The manufacturer or other parties wanting access to the information require the owner’s permission or court authorisation.
The Unpredictable Nature of Disruptive Changes in Technology
Significant improvements in the automobile like the automatic transmission, air conditioning, antilock brakes or advanced emissions controls took 15 to 30 years to affect the majority of vehicles on the road, with a relatively predictable pattern of adoption. This is largely a consequence of the relatively long vehicle life cycle, now at about 15 years, and the high value of investment in the current fleet.
However projections based on annual year-over-year fleet penetration may not apply to automated vehicles according to some experts. Currently, projections for the introduction of practical fully automated vehicles on public roadways range from as little as three years to more than 20 years, and changes to behavior and usage patterns involve many imponderables. The timing for introduction and implementation of this new technology is difficult to predict reliability because it is considered a disruptive game-changer.
Widespread adoption of automated vehicles is relatively certain, but all of the impacts of this technology on complex economic and social networks, and whether the majority of privately owned vehicles will become truly “driverless” over the near to medium term cannot be predicted reliably.
The Road Ahead
I) Intelligent collision avoidance
Intelligent collision alerting and avoidance features that are based on supporting the primary human driver for brief periods – to avoid a collision for example, or to warn of a conflict with another road user like a cyclist or pedestrian, have great promise and are quickly making their way into new vehicles. Volvo has established an objective of zero fatalities and injuries in its vehicles by 2020. This technology operates within our existing regulatory framework, and the human operator continues to have primary responsibility for the vehicle. Its implementation throughout the new vehicle fleet will occur over time. The reduction in collision and injury rates will occur gradually as the fleet is renewed.
II) Transition to driverless vehicles
The transition to driverless vehicles or personal mobility without a human operator, likely in an electric vehicle (another disruptive technology) will require adjustments to the legal and regulatory system, but these aren’t insurmountable. A revised definition of who has control over the vehicle will have to emerge. The owner of the driverless vehicle may be required to meet certain insurance requirements that cover the guidance system supplier in a similar manner to what lessees currently do when they insure a vehicle that belongs to the manufacturer.
III) Availability of safety data
Despite some very public lapses over the last decade, traditional automakers understand the current system in which they are responsible for safety defects and for making vehicle data available to government safety defect investigators. Outreach and education is needed for new entrants to the market who may not recognise existing relationships between themselves, government and their customers. Uber, for example, has defined the advent of driverless vehicles as an “existential” endpoint. But Uber requires public agencies it collaborates with to maintain secrecy around ridership and cost data that is normally readily available from public transit authorities. Is government ready to stare down companies that place “existential” roadblocks around sharing safety data related to their vehicle automation technology? Are existing Motor Vehicle Safety Act guarantees of access to safety information sufficient to cover situations where it does not reside in the vehicle but is in the cloud or on a server controlled by the vehicle automation provider? This issue needs to be resolved before fully automated vehicles are put into the hands of private users.
IV) Privacy issues around the ownership of onboard vehicle data
Privacy issues abound, as an automated vehicle will be collecting and transmitting information continuously while in operation. While automakers more or less accept that the vehicle owner has priority ownership of onboard vehicle data, technology companies appear to see things differently.
Current practice around the use of personal information is grounded in the concept of “consent.” APA’s experience is that consumers will almost always choose convenience and physical safety over their privacy rights. “Consent” expressed as a check-off to gain entry to a software application or technology is given quickly and with little consideration by the consumer – virtually nobody reads the fine print. A systemic and evolving approach to privacy issues is needed, within a hierarchy that balances public health and safety, and concerns over criminal activity, versus marketing or economic uses of the information.
V) Unintended consequences
We must be ready to respond to the law of unintended consequences: i) A vehicle on autopilot should likely carry a visible warning (say a marker light) for the human drivers in proximity, and presumably will be equipped with a form of beacon for other automated vehicles. The robot vehicle may never run a yellow light, or fail to yield for a pedestrian, whereas human drivers engage in these behaviors frequently, and drivers of vehicles following them expect this. The consequence may be a rash of autonomous vehicles being hit from behind, and other not at-fault collisions. This just shifts the locus of responsibility for collisions rather than reducing their frequency. ii) Drivers are already unsure about how to integrate current collision avoidance technologies into their existing habits, for example not being sure about how to combine three mirrors, the rear window, backup camera screen and audible warnings when reversing a vehicle. More of this confusion is imminent in the years ahead; Canada needs to be actively studying the human factors and commit to public education and other corrective measures if required.
VI) Transport Canada to build capacity
Transport Canada needs to “tool up” for the new era. The APA recommends that Transport Canada create a team of approximately three to five people with the appropriate expertise and resources , including engineers with the required computer coding and other specialties. This team would:
- Conduct or participate in multi-country testing of partially autonomous and autonomous vehicles;
- Engage in outreach and education for new entrants like Apple, Google and Uber, as well as existing automakers;
- Develop in-house defect investigation expertise with respect to partially autonomous and autonomous vehicles;
- Consider reviving a National Public Safety Organisations expert committee to deal with emerging issues affecting autonomous vehicles;
- Engage in timely and robust dialogue with the automakers and vehicle automation companies;
- When necessary, ensure that vehicle regulation is up to date.
VII) Federal-provincial-territorial collaboration
The provinces and territories and the federal government may wish to coordinate their activities, to avoid a piecemeal implementation of driverless technology. It is already becoming clear that a mixed fleet of human and robot drivers may not co-exist easily. In the early days of motoring, permission to drive on public roads was a municipal matter not really regulated outside incorporated areas with the exception of Great Britain where a person had to be walking in front of a motor vehicle with a red flag at all times. Eventually, larger authorities like provinces and states in Europe and North America stepped in. We are currently witnessing a similar phenomenon, with municipalities once again being the original innovators for the permitting of driverless vehicles. A more uniform and coordinated approach to standards-making and guidance based on best practice is necessary.
--
About the Automobile Protection Association (APA)
The Automobile Protection Association (APA) is a non-profit public interest organisation that promotes the consumer interest in the marketplace for automobiles and their related products and services. The APA was founded in 1969 by two people from Ralph Nader’s public interest group in Washington D.C. The Association has offices in Montreal and Toronto with public information hotlines and a website; we are a clearing house and advocate for consumers with auto related questions and complaints. The APA has counsel on retainer and is active as a plaintiff or in support of consumers in small claims courts and class actions. The APA performs unique field audits of vehicle retailers to evaluate regulatory compliance and their business practices. The Association also has a comprehensive new vehicle testing program conducted in partnership with Protégez-Vous, Quebec’s premier French language consumer publication. The APA is a small group that leverages its resources thanks to formal and informal arrangements with auto industry experts, many of whom help the Association discreetly and with no compensation.
|
