Technology Transfer

Prediction Technology of Lane Change
  • Year:2021
  • Project:DOIT Program
  • Category:Imaged assist technology
  • Technique Features:

    Accidents occur frequently when drivers fail to notice another vehicle in their blind spot during lane change. This technology provides a method to predict lane change intention of other vehicles in front of host vehicle via AI model with learning ability, and the host vehicle receives real-time warnings in advance as collision risk is high.

  • Specification:

    1.Host velocity < 50 KPH 2.Detection range / update rate: 40 meters for front and 4 meters for left and right with update rate> 10 3.Obstacle targets: Vehicles, motorcycles/bikes 4.Obstacle Velocity < 50 KPH 5.Driver behavior : i. Lane follow; ii. Lane change 6.Predict Accuracy > 85%

  • Applications:

    Autonomous System; Intelligent Driver Safety System.

  • Benefit:

    The technology provides a method to predict the vehicle intention of lane changing based on the AI model which had learned driving behavior of other vehicles, and the host vehicle receives real-time warnings in advance in dangerous situations based on our system.

  • Contact:

    rdservice@artc.org.tw