The Reality of Autonomy: Where Are We With Self-Driving Cars?

When will self-driving cars be the “new normal”? We seem to have turned a corner, with autonomous vehicles making headlines regularly for a variety of reasons, but it appears that there may be a long road ahead before we’re all completing sudoku puzzles on our morning commutes.

What Are the Hurdles and Setbacks?

One of the biggest barriers, although certainly not the most urgent, is the lack of consumer confidence in driverless cars. A survey conducted by AAA in spring 2018 revealed 73 percent of American adults would be too wary to ride in a fully autonomous vehicle. That number actually rose 10 percent over the previous year.

However, if the public is warier today of driverless cars, it’s probably because the concept is known to more people today than ever before. And that brings us to the next hurdle. When it comes to autonomous vehicles, the phrase “no news is good news” doesn’t seem to apply.

If anything, the presence of driverless cars on our streets is pushing the news media’s hunger for sensationalism and tragedy to its worst extremes — and it has the potential to worsen, rather than improve, the public’s perception of this technology.

In other words, if you simply tally up the number of times a Google, Uber or Tesla vehicle has been involved in a traffic accident, things look grim. But a closer look reveals that, predictably, we’re cherry-picking the worst outliers from the road test data. In-depth studies of traffic accidents reveal undeniable patterns in human behaviour and physical limitations that machines will not be constrained by. Simply put, humans cause most self-driving car accidents, as well as most other types of vehicle incidents, and probably always will.

There are two inconvenient realities about the state of self-driving cars right now:

  1. There are no cars currently available for purchase that can drive themselves fully autonomously.
  2. Despite this fact, many drivers remain confused about precursor technologies, including Lane-Keeping Assistance and Adaptive Cruise Control.

Adaptive Cruise Control helps drivers maintain steady and safe distances between themselves and other drivers. Lane Assistance alerts drivers when they stray outside the markings indicating other lanes or the edge of roads. They’re important “proof of concept” technologies and parts of the eventual apparatus that will help our cars drive themselves safely. But right now, they occupy a potentially confusing place in the world automobile market, with a surprisingly large and growing pool of customers who confuse these features with self-driving functionality.

What’s the Latest on Self-Driving Cars Breakthroughs?

The only question on most consumers’ minds is this: when will I be able to buy self-driving cars of my own? Not one with “lane assist” or a watered-down “autopilot” setting but one that’s actually autonomous?

According to automotive industry organisations and standards groups like the Institute of Electrical and Electronics Engineers (IEEE), driverless technology will have advanced to the point of being available as an upgrade on a wide variety of consumer-level vehicles by 2024. By 2044, those same groups say, driverless technology will have become mandatory on newly built vehicles rather than just a recommended factory upgrade.

[perfectpullquote align=”full” bordertop=”false” cite=”” link=”” color=”#8D3AF9″ class=”” size=””]Suggested Reading: Why Our Driving Habits Are A Reflection Of Our Civility[/perfectpullquote]

Clear lines in the sand, legally speaking, are some of the first requirements before this happens. Precedents are emerging which clearly define the automobile manufacturer as the liable party in the event of a collision or fatality. This shift in responsibility makes sense, given the shift in who, and what, is actually doing the driving. Fewer accidents overall should mean falling car insurance premiums for motorists, but the cost of keeping their cars and companies insured might mean automobile manufacturers pass the price of liability on to their customers anyway.

So the question is, what needs to happen in the meantime? Some of the major steps that need to be taken for any of this to be realised have already been taken, or are at least in progress. Here’s a breakdown of what’s changing as we speak, or will have to change in the near future, for these cars to catch on and change the world.

1. City-Based Testing Will Continue to Roll Out

Multiple companies are actively testing their own self-driving cars on American roads. Waymo is one of these, and currently operates in and around Silicon Valley, Mountain View, Los Altos, Palo Alto and Sunnyvale in California.

In fact, cities are taking the lead when it comes to approving tests for driverless cars on public roads. This is one place where local and state governments can make better decisions than the federal government when it comes to drawing up rules that reflect the needs and peculiarities of our cities and regions.

2. Countries Will Draw Up More Specific Regulations

At the national level, individual countries will have to come to their own conclusions about the “ethics” of self-driving car software. Germany is one country that’s ahead of the game. The German Ethics Commission on Automated Driving came to the conclusion that in an “unavoidable accident situation, any distinction between individuals based on personal features … is strictly prohibited.”

In other words, Germany is drawing a hard line in the sand when it comes to machine-based decision-making. If it finds itself faced with the infamous “trolley problem” and needs to minimise harm to bystanders, no vehicle may solve the dilemma by favouring one group of human beings over another based on superficial distinguishing characteristics.

3. Technology Will Increase the Number of Road Conditions Under Which Autonomous Cars Can Safely Operate

Arguably the biggest barrier to the more widespread rollout of autonomous vehicles is the technology itself. Most of the pathfinding tools aboard current attempts at autonomous vehicles are still inadequate for a number of common driving scenarios and conditions.

Rain and snow on the ground are the biggest culprits — they cause reflections that muddle the usefulness of cameras and lidar equipment and interfere with the car’s ability to “interpret” its immediate surroundings. Cameras don’t fare well while precipitation is still falling, either, or in heavy fog. GPS is useful but doesn’t have the precision necessary to help with obstacle avoidance.

Some proposed solutions involve burying radar equipment beneath roadways, but doing so would be hugely cost-prohibitive in any country that already struggles to meet its basic infrastructural needs. And keeping tomorrow’s vehicles dependent on in-road technology would be a blow to the “fully autonomous” vision most of us have in mind when we picture the future of transportation.

You might hear the phrase “Level 5 autonomy” used to describe this vision. This term describes a situation in which no human intervention is required for the vehicle to safely maintain its own functions, monitor its environment, react to obstacles and stimuli, adjust speed and change lanes as needed and keep itself on the road. This is the future that, according to the IEEE, we can expect as “standard” on automobiles by 2044. We have some work to do, but every passing day makes this prediction look more and more likely.