Tesla Model 3 dashboard in Autopilot testing with IIHS [CREDIT: IIHS]
Over the weekend, Tesla CEO Elon Musk tweeted that Tesla is testing new features on its Autopilot "self-driving" system, that will allow its cars to recognize stoplights and stop signs and negotiate traffic on roundabouts.
"Your Tesla will soon be able to go from your garage at home to parking at work with no driver input at all," Musk tweeted.
That sounds like something a lot of drivers would like. Musk has 23.6 million followers on Twitter around the world; 1,400 had commented on the tweet and 5,100 had retweeted it to others, as of yesterday afternoon.
Already testing traffic lights, stop signs & roundabouts in development software. Your Tesla will soon be able to go from your garage at home to parking at work with no driver input at all.— Elon Musk (@elonmusk) December 9, 2018
The tweet highlights how far along Tesla may be in developing its self-driving software and also how its drivers may become more reliant on its Autopilot system than Tesla itself says they should be.
Other companies testing self-driving software, such as Google's Waymo and GM's Cruise, already recognize stoplights and stop signs. They are testing only in limited areas, however, where their cars to try to encounter new obstacles one at a time to learn how to deal with each new variable.
Since cars aren't allowed on roads without a driver, each company has to apply for local permits to test cars driving themselves, especially when they begin testing with no-one in the car at all. Google recently announced that it would begin a commercial self-driving service where it is testing in Phoenix, but then didn't open the service up to the public, instead limiting it to a small group of pre-screened riders.
Tesla is taking a different approach. Since the company has owners driving its cars all over the world, and most of the cars have front and rear cameras, forward radar, and side proximity sensors—all of which can transmit driving data back to Tesla headquarters—the company is harvesting all this data from its drivers and running simulations using artificial intelligence software to improve its system's performance, updating the software and pushing improvements out to cars periodically, using other drivers' data to gradually improve Autopilot's driving.
Tesla CEO Elon Musk on
The company tweeted last month in time with the LA Auto Show, that drivers in its cars have covered 1 billion miles driving on Autopilot—giving the company a vast trove of data to improve the self-driving system.
Yet roads around the world aren't all alike, and it can be difficult for the company to integrate so many more driving scenarios so much more quickly, even with more real-world miles under their collective belts.
Tesla's cars have been involved in several high-profile crashes this year when driving on Autopilot, including one in which the car swerved into the damaged end of a highway divider and accelerated, killing its owner, and two in which cars under Autopilot control hit the back of stationary emergency vehicles and injured their drivers.
Each time, Tesla has responded by releasing a statement saying: "When using Autopilot, it is the driver’s responsibility to remain attentive to their surroundings and in control of the vehicle at all times. Tesla has always been clear that Autopilot doesn’t make the car impervious to all accidents, and Tesla goes to great lengths to provide clear instructions about what Autopilot is and is not, including by offering driver instructions when owners test drive and take delivery of their car, before drivers enable Autopilot and every single time they use Autopilot, as well as through the Owner’s Manual and Release Notes for software updates.”