Driverless Cars: Tomorrow’s Promise?

google guys in carWhile we await Governor Brown’s third veto of California’s safe-passing law AB 1371, we want to revisit an earlier piece of transportation-related legislation to which the good Governor did give his nod: SB 1298, the Google-friendly legislation that gave the green light (so to speak) to ‘driverless car’ operation in California. It was supported overwhelmingly in Sacramento (37-0 in the Senate and 74-2 in the Assembly) and was signed almost exactly one year ago by the Governor. What could it mean for cyclists?

Signed one year ago, SB 1298 ‘Autonomous Vehicles: Safety and Performance Requirements’ ostensibly regulates for safety this new innovation, but mainly it enables testing of driverless cars on public roads in the state. More importantly, the new law directs the California DVM to adopt regulations to govern the licensing and popular operation of driverless cars by January, 2015. This legislation was championed by Google – the primary proponent of ‘autonomous vehicles.’

Protecting cyclists from motorists is not so sexy, however. Advocacy organizations like the California Bicycle Coalition have labored to communicate our concerns to the Governor, but he’s vetoed safe-passing laws twice. And his silence this go-around is ominous. Why is it so difficult to codify a regulation that simply addresses one particularly negligent driver behavior? Perhaps because the California Highway Patrol and the Auto Club conspired to neuter it when they couldn’t kill it in the legislature?

With Governor Brown so reluctant to sign a simple safe-passing bill for the third time, we wondered Sacramento was in such a hurry to give its blessing to the driverless car – a quixotic, unproven technology with benefits a decade or more distant. One answer, not surprisingly, was that Google lobbied lawmakers for it. As far back as the 2009-10 session, it spread campaign largesse around and gave as much as $25,900 to Governor Brown (and his rival) during that election cycle, according to the Bay Citizen.

To be honest, we didn’t pay much attention when the Governor signed the bill in September of 2012. We were impressed by the technology (heck, even self-parking sounded exotic) but we were concerned about the ramifications. Isn’t dodging distracted or careless motorists enough? Do we need driverless cars too? If we are really better served by taking the motorist out of the driver’s seat entirely, then shouldn’t we focus our policy reforms to educate better drivers?

Now, we’re sympathetic to the argument that collision-avoidance mechanisms can save lives. We suppose the logical next step is to collision-proof the whole driving experience. But given policymaker haste, can the driverless car the panacea? After all, we’ll still have the same old roads uninformed by best practices we see around the world. We’ll still have problematic access to expensive health care, which leaves injured cyclists in the lurch after a collision. We’ll still have indifferent enforcement. And of course we’re stuck with the careless or negligent drivers that inspired the passage of AB 1371 in the first place.

For us the driverless car is no slam-dunk. If we can’t even get a safe-passing bill out of Sacramento, should we be satisfied with a few falling crumbs from the big loaf tossed Google’s way by the Governor? Let’s play devil’s advocate and look at the bright side – at least as advanced by driverless car proponents and prophets.

The Safety Argument

World Health road injuries infographicThe Bloomberg Road Safety Report found that 1.3 million lose their lives to collision every year – a cause of death projected to step up to the fifth-leading cause globally by 2030 as more new drivers find a place behind the wheel. The Foundation also tallied more than 20 million collision injuries annually, with 90% of them occurring in places less-economically privileged than California, where poor access to medical care, substandard facilities, and inadequate safety regulations all conspire to push victims into the mortality column.

These are also places largely without the safety measures we mostly take for granted. Their roads are sometimes poorly engineered and maintained. And the foundation found that nearly 85% of the world’s countries lack the laws necessary adjudicate responsibility for the increasing rates of traffic-related deaths and injuries.

Bloomberg traffic injuries table

Road traffic injuries are expected to climb as the leading cause of death globally.

When a major foundation focuses on the public health implications of traffic, just like Malaria or Polio, say, it’s time to take the high cost of motoring seriously.

Enter the ‘autonomous vehicle’ law. The California law says ‘safety’ right in its title, and indeed the bill text mentions (briefly) “significant potential safety, mobility, and commercial benefits.” Yet the bill text silent on the nature of potential safety gains. In fact, reading the following passage, one has to wonder if the new technology adds anything at all to existing safety-enhancing innovations:

An autonomous vehicle does not include a vehicle that is equipped with one or more collision avoidance systems, including, but not limited to, electronic blind spot assistance, automated emergency braking systems, park assist, adaptive cruise control, lane keep assist, lane departure warning, traffic jam and queuing assist, or other similar systems that enhance safety or provide driver assistance, but are not capable, collectively or singularly, of driving the vehicle without the active control or monitoring of a human operator. Section 1(2b)

So, do we need a car that drives itself? An insurance industry report provocatively-titled ‘A Scenario: The End of Auto Insurance’ suggests that the answer is yes. At least from the safety perspective. The ‘scenario’ predicts that new technologies including the driverless car will change the game. Namely, it will significantly reduce monetary losses from collisions so much that the proportion of the collected premium that covers property damages will drop to only one-quarter of what is collected by insurers from policyholders today.

Those new technologies include today’s collision avoidance features but add into the mix greater deployment of automated traffic enforcement, more pervasive use of telematics (e.g., collection of driver data) and, of course, the projected popular embrace of the driverless car.

The report envisions reduced overall auto insurance costs, but of course that means a changed insurance industry model: fewer collisions mean smaller losses and lower premiums and, presumably, smaller margins. In response to the prospect of the driverless car, proposals have emerged to socialize auto insurance costs.

But the report’s most salient prediction is the proliferation of policy-driven incentives to smooth sales of driverless cars. This handy timeline suggests that deployment may lag other policy initiatives, but that eventually Google’s pie-in-the-sky technology will find a ‘preferred’ place in consumer garages by 2023:

Driverless adoption timeline via Insurance

The Value of Data

google autonomous vehicleWhat if some cloud-computing entity outfitted a growing fleet of roving traffic data collection devices to aggregate a picture of driver behavior much like we aggregate data on speed and location to create a simple coarse-grained traffic map? But much, much more fine-grained and detailed of course.

As detailed in Google’s patent application (#20120083960) titled, ‘System And Method For Predicting Behaviors Of Detected Objects,’ a mountain of real-time road behavior data will be gathered by an ever-expanding fleet of driverless cars as part of its autonomous vehicle program.

driverless car infographic

Via the Economist

Moreover, Google wants to collect and analyze data by type of road-user. The idea is to refine an understanding about how certain road users behave and then predict how they might behave. The patent uses the example of trucks versus bicycles to highlight the differences and suggest how possible movements might need to be anticipated differently.

The broader idea is to learn from experiences on the road and then turn around and improve the autonomous vehicles’ response under known conditions but also react to unknown, albeit anticipated, events. The driverless cars will feed Google data for its algorithms and (theoretically) work toward fewer collisions & more efficient operation. The program would scale, too, as more cars that hit the road. We will know about real-world driving behavior. (A post on Google’s autonomous vehicle patents is forthcoming.)

Big data from the road also suggests that we might learn more about those drivers who  do remain behind the wheel. Today, concrete observations about road user behavior is  catch-as-catch can; our data are selective and always retrospective (for example court records). Plus, real-time driver data generated by automated enforcement tools remains siloed within jurisdictions. But those aren’t the only dots we’re not connecting today.

Couldn’t we also learn more about the extent to which particular road characteristics contribute to collisions? Like speeding, say? Today, if we have an hypothesis we could conduct a traffic study to test it. We’ve been making educated guesses about what kind of facilities don’t serve us well and incrementally introducing changes, but at a glacial pace. But that’s a far cry from real-time, crowd-sourced data. No wonder we can’t bend the death and injuries curves! Indeed there’s been no clear inflection point. Four decades after the bicycle renaissance in the 1970s, we’re still sifting through study findings to learn the net effect of bicycle lanes and road diets.

Looking ahead a decade, will driverless vehicles and their cluster of sensors provide something closer to real-time data and perhaps even feed back a change in prevailing speeds after a road diet, say? That kind of controlled test is long familiar to lab scientists and can be made available to departments of transportation too with the right data and analysis.  And that suggests both a new kind of toolbox and a new kind of traffic engineer.

The Environmental Promise

The environmental benefit claims on behalf of driverless cars is party premised on the safety argument: by reducing the incidence and severity of collisions, our cars can be made smaller and lighter and occupy a smaller environmental footprint. Some say that in wide use they could eventually reduce fuel consumption by a factor of ten.

Environmental claims are also premised on the efficiency argument: rationally-managed motoring will save time and fuel because our vehicles will know where to park and how to cue on the roadway. That reduces wind drag and increases capacity. And, taken to it’s extreme, such convoyed vehicles in ‘road trains’ could help to solve the oil crisis.

Environmentalists say that a tiny proportion (about 1%) of the fuel we burn for personal transportation actually transports the person. The rest moves mass and air or is lost in light and heat. With so much room for improvement, the of game-changing motoring innovations is vast. “Still, there is a very real possibility that within the next decade self-driving vehicles will begin to flip a century-old car regime on its head,” prophesizes the Yale environment blog.

“The rise of driverless cars would also affect the planning and layout of cities,” according to The Economist.

Assuming that autonomous vehicles make journeys quicker and use road space more efficiently, how should planners exploit the benefits of automation? On the one hand it would allow cities to get bigger, by reducing the time and stress associated with commuting. On the other, it could allow cities to become denser, by reducing the amount of space that needs to be dedicated to roads and parking. Alternatively, space allocated to roads in city centres could be used for bike lanes or parks.

The Driverless Car is No Substitute for a Safety Law

Renewed and reshaped cities perhaps, but the driverless car can expect some bumps in the road. Only two states in addition to California (Florida and Nevada) have enacted enabling legislation and the federal statutes don’t yet reflect the adjustments we would have to make to regulate these new unmanned urban terrestrial vehicles. The law is woefully unprepared, according to the Center for Internet and Society, as accords (like the 1949 Convention on Road Traffic have not kept pace globally.

Ethicists also question the wisdom of allowing the ‘Internet of things’ to handle the heavy machinery in crowded urban centers. Dystopians point to the future ‘hacked’ driverless car and raise the inevitability of robot cars going berserk on our roadways, for example. Even The Economist editorializes a call for “laws of robotics” to temper the growth of autonomous technologies like the driverless car.

Clearly we’re at the beginning of the story. We’re seeing boosters, hardware makers, and political patrons align. And of course the tools are rapidly evolving. Indeed if we listen to the driverless car boosters we might well think it safer to simply take the driver out of the driver’s seat! These devices may well deliver on the promise. But that’s tomorrow’s promise; cyclists still wait on the Governor’s signature to deliver at least one incremental step toward safer streets today: the right to be passed safely by a driver in a vehicle.

Resisting the Panopticon: Our Policies Change

Google's new networkIn yesterday’s post about Google’s announced service changes, titled A Panopticon Matures, we pointed out that starting March 1st the search giant will consolidate information about online activities across all proprietary platforms and fold it into real-name Google profiles. Known as ‘One policy, one Google experience,’ the new service changes are intended to create a social search experience that “does what you need, when you want it to.” Given those changes, we’re making some under-the-hood changes in how we interact with users. Hint: We’re not doing evil. Continue reading

On Google’s Terms, A Panopticon Matures

Google's new networkGoogle recently announced a change in how it collects information from search and online activity conducted through Google platforms and proprietary services like Youtube. While Google long collected IP addresses as a proxy for location, and also used cookies to identify return visitors, the search giant is going a big step further now: it will begin to marshal data from users’ activity data from across all of its services and then will attach it to a user account. All the better to know you and serve you ads. And so much more! Continue reading