SoylentNews is people
SoylentNews
from the I-want-to-drive-in-the-other-lane;-I-want-to-merge-like-humans-do dept.
Submitted via IRC for TheMightyBuzzard
In the field of self-driving cars, algorithms for controlling lane changes are an important topic of study. But most existing lane-change algorithms have one of two drawbacks: Either they rely on detailed statistical models of the driving environment, which are difficult to assemble and too complex to analyze on the fly; or they're so simple that they can lead to impractically conservative decisions, such as never changing lanes at all.
At the International Conference on Robotics and Automation tomorrow, researchers from MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) will present a new lane-change algorithm that splits the difference. It allows for more aggressive lane changes than the simple models do but relies only on immediate information about other vehicles' directions and velocities to make decisions.
[...] One standard way for autonomous vehicles to avoid collisions is to calculate buffer zones around the other vehicles in the environment. The buffer zones describe not only the vehicles' current positions but their likely future positions within some time frame. Planning lane changes then becomes a matter of simply staying out of other vehicles' buffer zones.
[...] With the MIT researchers' system, if the default buffer zones are leading to performance that's far worse than a human driver's, the system will compute new buffer zones on the fly — complete with proof of collision avoidance.
Let me know when someone finds an algorithm that can deal with unknown situations as intuitively as human beings can. Until then...
Source: http://news.mit.edu/2018/driverless-cars-change-lanes-like-human-drivers-0523
(Score: 3, Interesting) by VLM on Friday May 25 2018, @02:54PM (2 children)
My immediate reaction was a racing car course I took (nothing too exotic) where serious sober racers can trust each other to behave in a sane manner on a controlled-access track, so you can tailgate people at high speed without getting killed nearly as often as would happen on the road.
With a side dish of when I had a long work commute and sometimes I was on the interstate with civilian idiot drunk drivers and sometimes out with the rush hour professionals, there is a certain feel to the experience such that the commuting professionals pay much more attention and are vastly less unpredictable such that you could take certain liberties without dying. This is not necessarily extreme stuff, for example due to perceived fellow driver skill level I feel much safer at 25 MPH in ice conditions with commuter professionals than I do with Sunday drunk driver yahoos at 15 MPH. The amount of stupidity WTF you see on the road varies a bit with time of day. I wonder if the learning algos as mentioned are picking up on that. That guy looking out each window and mirror while wearing a suit and tie and its 5:00 pm I can inch a little closer to him than some moron playing with his radio and drinking god knows what out of a bottle as he weaves around at 2am on the way to a scheduled maintenance window, so that goofball gets the maximal size space cushion for safety.
Wait till the algos are discovered to be "racist" because certain skin color drivers are less predictable on the road than others... Or the algo learns "oh look there's a pot leaf bumper sticker aka that driver's reaction time will be measured in seconds not milliseconds", etc.
(Score: 2) by choose another one on Friday May 25 2018, @03:45PM (1 child)
No need to wait: http://www.dailymail.co.uk/sciencetech/article-5768539/Is-Amazons-facial-recognition-RACIST-Expert-says-distinguish-black-faces.html [dailymail.co.uk]
And don't get me started on voice recognition and scottish accent... (just google it)
(Score: 1, Funny) by Anonymous Coward on Saturday May 26 2018, @11:03AM
The scots do that on purpose. When they get drunk they talk like normal people.