The recent and shocking news that one of Uber’s self-driving vehicles struck and fatally wounded a pedestrian is making the rounds. In camera footage, it appears the vehicle did not slow down, and the safety driver — who wasn’t actually in control of the vehicle — looked down briefly before the accident. It’s disturbing, of course, but it also brings up an important question: Are we truly ready for driverless vehicles?
As a starting point, it’s also important to remember that — although tragic — the Uber fatality was a single incident out of thousands of hours that autonomous vehicles have already driven on public roads. That doesn’t minimize what happened, but it goes to show that fatal accidents are not the norm for this technology.
How Do Most Driverless Vehicles Work?
A plethora of auto manufacturers are currently working on self-driving technologies, from Ford and Tesla to BMW and Volvo. In fact, you’d be hard-pressed to find a car company, or for that matter a large tech firm, that’s not currently working on it.
For the most part, each system is different, offering a distinct set of features. One might be able to self-park, a second might be able to freely change lanes on the highway, while a third might have automatic braking and collision avoidance technology, or even fully autonomous driving modes.
Inside these vehicles are incredibly sophisticated systems, designed to capture and analyse vast amounts of information in order to make driving decisions in real time. They use a combination of big data, machine learning, and artificial intelligence to analyze current road conditions, and the surrounding environment, as well as to receive commands from passengers.
When it comes to the future of automated and driverless vehicles, big data is vital to the technology’s success. It’s not just about controlling the vehicles and allowing the systems to make split-second contextual decisions — it’s also about knowing the routes and travel areas. Each system must have a well-designed database of maps, charts, and details about the roads they are traveling on. This can also involve using geospatial information systems (GIS) to optimise travel routes, which requires a great deal of data management.
That’s why companies are spending a lot of time and money testing their vehicles before making them available to the general public. They need to make sure the systems work properly, of course, but they also need to map out our roadways in order to enable vehicles to better understand how to interact with the world at large.
It’s these big data and analytical systems that will be responsible for making split-second life-or-death decisions just before a critical event. In the case of Uber’s vehicle, its system failed to do what it was supposed to.
What Does This Mean for the Industry?
Right now, getting to the bottom of what happened is the most important step for Uber. In the long term, however, this incident just shows how difficult it will be to regulate and deal with this technology.
- Who is to blame when an autonomous vehicle acts (or fails to act) in a way that causes someone to suffer loss or harm? Can it really be the driver if they weren’t in control? Should it be the auto manufacturers, or perhaps the software developers who coded the AI and machine learning system?
- How can we tell whether the autonomous vehicles acted intentionally or negligently?
- If the injured person decides to seek compensation, who do they need to take to court?
These are things that will eventually be figured out, but let’s hope it happens sooner than later. Driverless vehicles are coming, and fast. Many auto companies predict that they will have driverless technologies operating on real roadways by 2020.
The Future of Autonomous Cars
At this stage, driverless vehicles are probably not safer than good old human piloted ones. It’s true that one Uber fatality seems to compare favorably with the 40,000 people killed on American roads each year. However, this statistic is misleading. The correct way to assess safety is to look at deaths per miles travelled. Since Americans drive trillions of miles per year, and driverless vehicles have only been test-driven a few tens of millions of miles, the human drivers are still certainly superior, for now.
When driverless cars do finally hit the roadways, they will probably still be less safe than the human alternative. However, early automobiles in the 19th century were also less safe than horse drawn carriages. We might expect that driverless tech, which is powered by software, will continue to be improved until its safety record makes a Volvo look dangerous.
So, what’s in store for the future? Where is this technology truly headed?
The answer is that every type of vehicle, from public transport to freight, may soon be automated. Tesla, among others, is already working on self-driving freight trucks. This is a sensible place for them to start, because there is a clear value proposition. Trucking is a multi-billion dollar industry and drivers represent the second largest operating cost. Median truck driver salaries are more than $40,000 per year, and since truckers are human they also need to take rest breaks, which increases long haul delivery times.
Within the next few years, we could be looking at autonomous vehicles dominating our roadways. Frightening or not, that’s the reality we’re in.
Nathan Sykes enjoys writing about technology and business online. He is the founder of Finding an Outlet, a tech blog he runs out of Pittsburgh, PA.
Image: Pexels