The application discloses a lane line optimization target
screening method with state
hysteresis, and belongs to the technical field of intelligent driving and vehicle longitudinal control. The method first acquires lane line parameters and vehicle motion parameters through a camera and a vehicle body sensor, calculates a lane center line, a short-
time trajectory of the vehicle and a lateral error, and extracts a first derivative of the lane center line; then, a state
machine containing six behavior states is used to realize accurate
behavior recognition in combination with a lateral error and a first derivative threshold constraint, so that road bending and real lane changing misjudgment are avoided; when the behavior state is switched from lane changing having crossed a line to centering / offset keeping, a preview deviation
limiter timer is started, and the sudden change of the preview lateral error is suppressed within a timing window; and finally, a path is screened based on an optimized lane center
line fitting, and target screening is completed by adaptively scaling a
channel width. The application improves the stability and accuracy of target screening under complex working conditions, reduces unnecessary acceleration and deceleration, and enhances driving comfort and
system robustness.