End-to-end vehicle trajectory prediction method based on machine vision

• Main Authors: ZHANG BINGLI, SHE YAFEI
• Format: Patent
• Language: Chinese; English
• Published: 14.07.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention discloses an end-to-end vehicle trajectory prediction method based on machine vision, and the method comprises the steps: 1, a feature extraction network: carrying out the feature extraction of a plurality of monocular camera images through a convolutional neural network; 2, a motion prediction network is used for predicting motion tracks of surrounding vehicles at future moments by introducing two GRU models; according to the method, the problems of feature loss and error accumulation in the motion trail of the surrounding vehicle in the future within a short time in automatic driving can be solved, so that the prediction accuracy can be improved, and the driving safety can be guaranteed. 本发明公开了一种基于机器视觉的端到端车辆轨迹预测方法，包括：1.特征提取网络，利用卷积神经网络对多个单目相机图像进行特征提取；2.运动预测网络，通过引入2个GRU模型来预测周边车辆未来时刻的运动轨迹。本发明能解决在自动驾驶中周围车辆短时间内未来时刻运动轨迹中的避免特征丢失和误差累积的问题，从而能提升预测准确度并保障行车安全性。