Method for solving problem that vehicle prediction model performance is limited by perceptual data performance

• Main Authors: QIN XUEFANG, ZHANG HANCHI, XU YUNENG
• Format: Patent
• Language: Chinese; English
• Published: 01.04.2022
• Subjects: CONTROLLING; PHYSICS; REGULATING; SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES

The invention provides a method for solving the problem that the performance of a vehicle prediction model is limited by the performance of perception data, and the method comprises the steps: selecting a predicted vehicle in a set range, and obtaining a vehicle prediction track of the predicted vehicle outputted by a deep neural network model; according to the perception data of the predicted vehicle, extracting center line data of a lane where the predicted vehicle is located currently from a map database; correcting the predicted vehicle track, namely selecting a plurality of coordinate points on the predicted vehicle track as predicted vehicle track points, adjusting the distance between the predicted vehicle and the lane center line according to the deviation degree of the predicted vehicle relative to the lane center line, and obtaining a corrected predicted track of the predicted vehicle; and the steps are repeated, and the corrected prediction tracks of all the vehicles in the set range are obtained.