A new algorithm developed by the Massachusetts Institute of Technology's (MIT) Computer Science and Artificial Intelligence Laboratory (CSAIL) could improve the way driverless vehicles change lanes. Compared to existing models, the new algorithm allows for more aggressive lane changes while avoiding a detailed statistical model that cannot be analyzed on the fly. If default buffer zones are leading to performance that is significantly worse than a human driver's, the system will compute new buffer zones on the fly, with proof of collision avoidance.
The researchers used a Gaussian distribution to represent the current position of a car. Based on estimates of the car's direction and velocity, the system generates a logistic function that, when multiplied by the Gaussian distribution, skews the distribution in the direction of the car's movement. This skewed distribution defines the vehicle's new buffer zone, using only a few equation variables so that the system can evaluate it on the fly.
In a simulation with up to 16 autonomous cars driving among several hundred other vehicles, the autonomous vehicles ran the algorithm in parallel without conflict or collisions.
From MIT News
View Full Article
Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA