By: Michael Schmall, Vice President; Sara Beauchaine, Marketing Associate
Google, Apple, Tesla, and leading car manufacturers around the world are well on their way to producing autonomous vehicles, with the expectation that some of these cars will be on dealership lots within the next 10 years.
Autonomous qualities are already being incorporated in vehicles, and, for example, can be seen in Tesla’s products as “Autopilot”. This technology utilizes multiple sensors, radar, cameras, and sonar to pick up on road lane lines and other vehicles, allowing the car to essentially drive itself on expressways. This technology requires the driver to—at the very least—keep one finger on the wheel.
With advancements in technology such as “Autopilot” leading the way, a logical question looms on the horizon: Will artificial intelligence eventually come into the equation of manufacturing vehicles to be completely autonomous?
Artificial intelligence has been around for years, and within reach of consumers. It can be seen in Apple’s Siri, GPS units, and many other devices. Automotive companies like Toyota are now investing heavily in artificial intelligence, and technology tinkerers are attempting to bolster autonomous cars’ driving abilities with this technology.
One of these tinkerers is George Hotz. 26-year-old Hotz not only built a self-driving car in his garage by himself, but is also programming it with artificial intelligence—rather than manually coded technology. According to a recent article, Hotz revealed that incorporating artificial intelligence software in self-driving cars will help avoid the common roadblocks currently experienced by other autonomous test vehicles.
Some of the common obstacles currently facing manually coded autonomous technology, like the Google car, include:
- Human hand signals (such as those from an officer directing traffic)
- Small animals crossing the road (larger living creatures, such as deer and jaywalkers, are sensed by the car)
- Bad weather (snow, fog, and splashing ground water interferes with how the car receives and interprets information)
- Areas with no cell signal (operation ceases without a connection to a cell signal, as critical access to GPS maps is cut off)
Hotz believes that some of these obstacles can be avoided by employing “deep-learning techniques in autonomous technology”.
Artificial intelligence software, like what Hotz is producing, watches the driver’s behavior. It learns as it observes how certain situations are typically handled by the human driver, and then makes its own decisions by mimicking these learned behaviors and actions when operating autonomously.
If artificial intelligence can truly be utilized for autonomous vehicles, we may soon find solutions to some of the issues that need to be resolved before these cars can be introduced to the market. There is a possibility that soon, cars might be watching you, too!