The invention provides a multi-modal vehicle trajectory prediction method based on a hierarchical order network in the field of unmanned driving trajectory prediction. The method comprises the following steps: acquiring and converting historical trajectory information of traffic participants and lane information of a target vehicle; based on historical track information of a target vehicle and lane information of the target vehicle, a hierarchical order network is designed to perform spatial-temporal feature and scene information extraction on a local area and perform message transmission on a global area; fusing the local region spatial-temporal features and the global region spatial-temporal features to obtain future trajectory fusion features, and obtaining multi-modal trajectory prediction results of surrounding vehicles and possibilities thereof based on the future trajectory fusion features; designing a multi-task loss function according to the multi-modal trajectory prediction result of the surrounding vehicles and the possibility to calculate model loss, and optimizing and updating a weight coefficient through back propagation until the model is converged; the running speed of the multi-modal vehicle trajectory prediction method based on the hierarchical order network on a computing platform meets the real-time requirement.
本发明提供无人驾驶轨迹预测领域一种基于分层次序网络的多模态车辆轨迹预测方法,包括:获取并变换交通参与者的历史轨迹信息以及目标车辆的车道信息;基于目标车辆的历史轨迹信息和目标车辆的车道信息设计了一种分层次序网络对局部区域进行时空特征和场景信息提取和全局区域进行消息传递;将局部区域时空特征和全局区域时空特征进行融合获得未来轨迹融合特征,基于未来轨迹融合特征获得周围车辆多模态的轨迹预测结果以及它们的可能性;将周围车辆多模态的轨迹预测结果以及可能性设计多任务损失函数计算模型损失,通过反向传播优化并更新权重系数直至模型收敛;基于分层次序网络的多模态车辆轨迹预测方法在计算平台上的运行速度达到实时性要求。
Multi-modal vehicle trajectory prediction method based on hierarchical order network
一种基于分层次序网络的多模态车辆轨迹预测方法
2023-08-01
Patent
Electronic Resource
Chinese
IPC: | B60W CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION , Gemeinsame Steuerung oder Regelung von Fahrzeug-Unteraggregaten verschiedenen Typs oder verschiedener Funktion / G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen |
Interactive vehicle multi-modal trajectory prediction method based on GRU-GCN
European Patent Office | 2023
|Vehicle multi-modal trajectory prediction method based on semi-supervised model
European Patent Office | 2023
|