Abstract Three algorithms to detect driver’s lane-changing maneuvers are put forward there. Feature sets in different dimensions, several time windows and classifiers based on different theories are analyzed. Simulation result shows that the parameters including land departure amount, vehicle acceleration, steering angle, steering angle velocity, and steering torque can validly reflect the vehicle running status. The Random Forest (RF) Classifier based on decision-making tree theory presents the best detection result, where the lane-changing detection result reaches 94.92 % and the lane-changing maneuver can be detected 0.04 s after the onset of steering and 2.37 s before the vehicle totally crosses the desired lane line.
Drivers’ Lane-Changing Maneuvers Detection in Highway
Man-Machine-Environment System Engineering ; 1 ; 21-29
Lecture Notes in Electrical Engineering ; 406 , 1
01.01.2016
9 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Random forest , Automotive engineering , Lane-changing maneuver detection , Support vector machine , Driving simulation , K nearest neighbor Artificial Intelligence (incl. Robotics) , Aerospace Technology and Astronautics , Human Physiology , Engineering Design , Engineering , Quality Control, Reliability, Safety and Risk , Robotics and Automation
Drivers' Lane-Changing Maneuvers Detection in Highway
British Library Conference Proceedings | 2016
|Lane Changing Maneuvers and Risk Analysis in Vehicular Network
Springer Verlag | 2024
|Modeling the impacts of mandatory and discretionary lane-changing maneuvers
Online Contents | 2016
|