The Critical Difference Between Autonomous And Semi-Autonomous

February 07, 2018

Autonomous cars could be the solution to so many road-related conundrums. Now that several driverless vehicle fleets have been deployed in cities around the world, there is plenty of data to show that robots are much better than people at piloting two-ton machines at high speeds.

In more than 2 million miles on U.S. streets, Waymo (Google’s autonomous vehicle program) has only been labelled at-fault for one collision; all other incidents with Waymo cars were caused by human-driven vehicles nearby. That’s 10 times lower than the average accident rate of the safest demographic of human drivers. Plus, autonomous cars could reduce emissions, eliminate traffic jams, and disrupt the market for car ownership in a manner beneficial to average consumers.

Semi-autonomous cars are not autonomous cars.

Because autonomous cars provide all the above-listed positives, interest in robot features is skyrocketing. As a result, many carmakers are building driver assist programs into their non-autonomous vehicles and labelling them semi-autonomous. Unfortunately, that is leading many drivers to believe that their cars are smart enough to drive themselves. As a result, collisions involving semi-autonomous vehicles on “autopilot” are climbing sharply.

If we want driverless vehicles to succeed, it is important to make the line between autonomous and semi-autonomous right now.

How Semi-Autonomous Works

Like autonomous cars, semi-autonomous functionality relies on a mixture of radar, sonar and cameras to understand the vehicle’s surroundings and send the appropriate messages to the driver.

  • Radar. Pulses of high-frequency electromagnetic waves moving as much as 160 meters (nearly 525 feet) identify the distance to objects nearby.
  • Sonar. Super-low soundwaves reach out 8 meters (about 26 feet) in every direction to assist in determining object location, distance, and speed.
  • Forward-looking side cameras. Simple cameras, positioned for 90-degree exposure on both sides, designed to detect oncoming vehicles up to 60 meters (about 196 feet) away.
  • Main forward cameras. Several cameras of various lens angles to detect details of surroundings, such as traffic lights and signs, as well as long-range analysis at high speed.
  • Rear-view cameras. Cameras for driver use in reversing and parking as well as detecting approaching vehicles up to 50 meters (164 feet) away. 

To the average driver, the technology involved in semi-autonomous vehicles seems identical to that of autonomous vehicles — but there is one critical difference: the computer. Truly driverless cars include incredibly powerful computers that take in the information gathered by the radar, sonar, and cameras and use it to make calculated decisions. In semi-autonomous cars, that tech is replaced with human brains — and if those human brains aren’t doing the necessary work, there’s trouble.

The Dangers of Semi-Autonomous Cars

It seems like a good idea to give human drivers some help behind the wheel. Our eyes can’t see everything, but continuous radar and sonar can; if a hazard creeps up that we can’t see, the car can alert us to the danger and assist us in navigating away. Semi-autonomous features are intended to provide humans with control while reducing human error.

Unfortunately, that’s not what humans are using this tech for. Research on driver behavior behind the wheel of semi-autonomous cars indicates that drivers, instead, are trusting the inadequate tech with driving responsibilities while they use their smartphones, solve mental puzzles, or otherwise fail to pay sufficient attention to the road.

In an autonomous car, this isn’t an issue. The car is designed to allow drivers to become passengers, giving everyone in the car the freedom to enjoy their favorite distractions. However, in a semi-autonomous car, the vehicle’s sensors do not have the power to react fully when danger is sensed. A distracted driver is slow to take action, and in high-speed situations, even semi-seconds matter.

The Insurance Institute for Highway Safety (IIHS) has tested several semi-autonomous vehicles and discovered programming mistakes that endanger everyone on the road. Waymo recognized the risks of permitting semi-autonomous features in their vehicles and decided to eliminate the option altogether. Some manufacturers have listened to IIHS and installed safeguards to encourage driver concertation; for example, Tesla’s Autopilot disengages if a driver’s hands leave the wheels, and GM vehicles track drivers’ eyes to ensure they remain on the road. However, almost all car manufacturers continue to release cars with semi-automation — and accidents related to the tech have continued to climb.

Roads filled with autonomous vehicles are still a decade or more away, but semi-autonomous cars are already serious risks to anyone on the road. While we should look forward to a driverless future, we should be wary of the real, semi-driverless present.

Jackie is a content coordinator and contributor who creates quality articles for topics like technology, business and education. She studied business management and is continually building positive relationships with the internet community.

Overlay Init

Curated By Logo