What Is a Self-Driving, or Autonomous, Vehicle?

can be reached at bev_b@outlook.com
can be reached at bev_b@outlook.com
Share on facebook
Share on twitter
Share on pocket

Even though some days it seems we’re closer to colonizing Mars than we are to traveling in flying cars, the idea of self-driving vehicles sometimes feels just as far away. Truth is, we’ve been working on Automated Driving Systems (ADS) for nearly a century.

Research on the matter has been conducted as early as the 1920s. In 1926, for example, a “phantom auto” drove through the streets of Milwaukee. Although functioning more like a life-sized, radio-controlled car with its operator following in a nearby vehicle, the car did start up, drive, and stop without a human driver behind the wheel.

Fast forward to the 1980s, when the Defense Advanced Research Projects Agency (DARPA) funded the autonomous land vehicle (ALV) project. Part of the U.S. Department of Defense, DARPA utilizes new research from institutes like Carnegie Mellon University’s Navlab and the Environmental Research Institute of Michigan (ERIM). The success of the project led to the first time lidar, artificial intelligence (AI), and autonomous robot controls were demonstrated on a vehicle.

More Than Five Senses

Uber Self Driving Autonomous Volvo XC90 with Sensors

The road to autonomy is paved with good sensors. From radars to lasers, the more information that is captured, the better a vehicle can “see” its surrounding environment.

Today’s ADS still rely heavily on these technologies as well as other types of sensors, even the “cloud.” Vehicles with semi-automated systems can use the following:

  • Ultrasonic sensors: Taking a lead from how bats “see” in the dark, these sensors use sound waves to detect an object based on its echo signature. They are used in low-speed and close-proximity detection.
  • Image sensors: Doing their best to imitate human eyesight, multiple cameras gather images to generate a picture of a vehicle’s surroundings. These cameras also can detect colors and fonts (i.e., to read road signs), and stereo cameras can create 3D images.
  • Radar sensors: Just like in oceanic and aviation navigation, Radio Detection and Ranging (radar) utilizes electromagnetic waves to detect obstacles, including their distance and speed (if approaching).
  • Lidar sensors: Light Detection and Ranging (lidar) uses a non-visible laser to scan the outside environment to create a 3D image. When combined with image sensors, lidar can distinguish between a vehicle and a pedestrian.
  • Cloud: All the information gathered by the various sensors must be stored somewhere, and a massive onboard hard drive doesn’t make sense. With the cloud, collected data also is constantly updated to provide real-time intelligence and mapping, just like how navigation systems can detect upcoming traffic or road closures.

Six Degrees of Automation

Man driving a Cadillac with Super Cruise

Many new vehicles have automated driver-assistance features but based on classification levels, none are remotely considered autonomous. Shown here is Cadillac’s Super Cruise system in action, one of the most advanced semi-autonomous driving technologies available at the start of 2019. (Cadillac)

Due to the mix-and-match packaging of safety features and driver-assistance technologies, the Society of Automobile Engineers (SAE) International developed a classification system (that governing bodies have also adopted) to readily identify a vehicle’s automated capabilities and limits. There are six driving modes:

  • Level 0 – No automation: There may be some automated systems like cruise control or proximity alerts, but they will not intervene with driving. A human driver is the sole vehicle operator.
  • Level 1 – Driver assistance: Vehicles fall under this category if they are equipped with at least one advanced driver-assistance system (ADAS), like adaptive cruise control or lane-keeping assist. And although a Level 1 vehicle can maintain speed and direction, driver supervision is still required. Many of today’s ADAS qualify as Level 1.
  • Level 2 – Partial automation: Vehicles must be equipped with at least two ADAS that can take full control of the steering, acceleration, and braking. However, a human driver must still monitor the vehicle and be prepared to intervene at any time. Example systems include Cadillac Super Cruise, Nissan ProPilot Assist, and Tesla Autopilot.
  • Level 3 – Conditional automation: The vehicle is capable of full dynamic driving, but a human driver must be able to regain operational control if needed. The redesigned 2019 Audi A8 is the first vehicle to claim Level 3 capability through its Traffic Jam Pilot system.
  • Level 4 – High automation: This vehicle is capable of complete vehicle operation and can drive a route with or without a driver. Some possible constraints could be geofencing and speed limits. There are currently no Level 4 production vehicles available.
  • Level 5 – Full automation: This vehicle is #lifegoals. Expected to handle all driving duties regardless of roads and weather conditions, Level 5 vehicles would be replete of a steering wheel, foot pedals, and gear shifters.

Is “Automated vs. Autonomous” the New “Potato vs. Potahtoh?”


Volvo Pilot Assist driver assist technology
There is still confusion between what today’s cars are capable of and what autonomy really means. It would behoove policymakers and marketers to use a common terminology when discussing these new technologies. (Volvo)

There currently is no standard terminology with regard to self-driving vehicle technology, but government policies, as well as the SAE International, use automated, like in ADS, as opposed to autonomous, even though the latter is more widespread in usage.

Neither is technically incorrect, but the lack of official guidelines has created confusion – sometimes with deadly results. As driver-assistance features continue to be bundled and receive some “pilot”-related nomenclature in marketing campaigns, consumer expectations only grow higher than what the systems can actually achieve. It’s a line in the sand, but a dangerous one.

From a linguistics standpoint, autonomous means independent while automated means controlled by a machine. Following this logic, safety features such as adaptive cruise control, lane-keeping assist, and automatic emergency braking are automated systems.

Although they may engage (braking) or disengage (cruise control) on their own, the features do not independently think on their own. Simply put, a truly autonomous vehicle is one where a passenger can request to be driven to one location, but (if slightly comical) the vehicle decides to take them elsewhere.

A hands-off, fully automated vehicle is not yet available to the public. However, General Motors has been vocal about having a Level 5 vehicle ready and on the road in 2019. In reality, experts believe true automation will not occur for another vehicle generation or two, which means another decade of waiting at a minimum.


Even so, policymakers have already introduced legislation regarding ADS and driving on public roads. In 2011, Nevada became the first state to authorize the use of autonomous vehicles, and since 2012 at least 41 states and the District of Columbia have considered similar legislation.

Regardless of when they finally arrive, Level 5 fully autonomous vehicles are indeed coming. With improved infrastructure and better public understanding of the technologies, the roads could be the safest they’ve ever been.

About the Author

can be reached at bev_b@outlook.com
Close Menu