What is an Autonomous Vehicle?
While it might be easy to take for granted what an autonomous vehicle is – ostensibly, a vehicle that drives itself – this a question that has become extremely pertinent. Defining and delimiting what counts as an autonomous vehicle has become essential to maintaining the safety and regulation of our roadways and the automotive industry. Let's start off by taking a look at the history of the idea of automation, and automated vehicles.
Automation as a notion and ideal undoubtedly first came to prominence during the industrial revolution. Factories and cities were sites where the previously scattered and slow work of production could be simplified and concentrated. Whereas individual craftsmen would need years of training and weeks of effort to produce just a few goods, factory assembly lines made it possible for relatively unskilled laborers, trained quickly to carry out straightforward tasks, to make those same goods at a similar level of quality, in a fraction of the time.
In fact, attempts to automate motor vehicles did not come so long after industrialization as one might think! As early as the 1920s engineers would begin to experiment with new technologies and find ways to render a car driverless, or less driver dependent. Of course, due to the primitive nature of these technologies, nothing produced would be considered road-ready (able to share space with human drivers), though there was some success in automating certain vehicles enough to have them navigate heavily constrained courses and tracks already by the mid twentieth century.
Of course, automation does not mean the same thing as being autonomous. However, I believe that this history shows a certain continuum between the two concepts. Being autonomous is the far endpoint of automation: the automotive – it's very name constituting a hint – was itself a stage of automation past the horse and carriage. Automation is the progressive march away from our own sustained involvement with the course of a task. 'Autonomous' is a term defined as self-governing, and thus does represent the logical extreme of automation: the absence of us from the task carried out by a machine.
A classification system created by the Society of Automotive Engineers (SAE) also reflects this continuum. SAE's rankings consist of six levels, and these levels are the de facto standard that is employed in public discussions of automation and autonomous vehicles. You will often see, for instance, figures like Elon Musk of Tesla Motors stating that SAE Level 5 is on the horizon for their vehicles (a claim regularly met with great scrutiny). These levels range from zero to five and are distinguished on the basis of pre-existing technologies and features planned to be built into cars. Level zero is characterized as minimal or momentary automated responsiveness. Level zero features are those like proximity warnings or an automated switch to all wheel drive in the case of lost traction. Level five, meanwhile, is endearingly entitled 'steering wheel optional' – expressing the total absence of driver input.
2022 is certainly the closest we've been to a future populated with autonomous vehicles. Tesla is more than a year into their controversially implemented autopilot beta, meanwhile companies like Waymo and Google are using remote-supervised self-driving cars to taxi customers and map the world, respectively. While these developments give the impression of autonomy, due to the strict stipulations employed by the SAE in their ranking systems nothing yet in development exceeds level four, meanwhile nothing presently available to consumers exceeds level two. This begs one to question how much we ought to abide by the SAE's technical ranking system. For instance, if Waymo can drive you from point A to B with no operator in sight, there must be some sense in which that taxi is autonomous. So, the limitation of SAE ranking is evident.
At the moment, two brands are heading the 'autonomous' vehicle market in the US – Tesla and GMC. Within the next year or so we will start to see foreign manufacturers entering the fray, with Honda planning to compete aggressively. While Tesla is known for their self-driving cars, perhaps because of Musk's techno-optimism and fame, or the brand's ubiquity, GMC's 'Super Cruise' system is actually considered a much stronger competitor – at least according to expert opinion.
As stated earlier, Tesla's autopilot mode is an SAE level two technology. This is true of its stock 'driver assist' feature set, which can, for instance, handle lane changes and parking, or be called to the driver's location, within line-of-sight. However, Tesla has technically released a beta (available to certain drivers and employees under restricted circumstances) which is considered SAE level four. This is possible only by acting in a legal grey area and using methods of testing which have long been abandoned by other manufacturers over safety concerns.
GMC's Super Cruise takes what one might consider the opposite approach of Tesla. Tesla's drive assist and autopilot have suggested conditions, but no hard limits on when the technology can be engaged. Super Cruise errs on the side of safety and assured performance by only being possible to engage on predetermined highways. In spite of this stark limitation, the experience is universally considered to far exceed that of Tesla.
For someone not truly acquainted with the state of 'autonomous' vehicles or driving automation, this blog might come as a surprise. Indeed, popular discourse has somewhat obscured the true state of 'smart' driving. The future of true autonomous vehicles is more distant than Elon Musk would have you think, and substantial legislative and technical hurdles suggest that truly autonomous rides are not quite on the horizon.
That said, the automation that is featured in Tesla and GMC vehicles will make for a less stressful, less tiring drive. As the technology refines, more roadways will be viable for drive automation, and the roads will become safer. While your colleagues Tesla may spark envy, it might cool the burn a little to know that having one doesn't mean you're quite in charge of your own robotic chauffeur.