The invention relates to a vehicle multi-modal trajectory prediction method based on a semi-supervised model, and the method comprises the steps: obtaining the future possible intention of a vehicle, a possible driving road and a central line based on the road map information, and carrying out the coding of the map information of the road around a target vehicle and the historical trajectories of other vehicles around the target vehicle, and then a semi-supervised model is used to generate a pseudo label for a future possible target vehicle trajectory mode, and the pseudo label is used to carry out supervised learning on prediction trajectories in different modes to realize multi-modal prediction of the vehicle trajectory. Compared with the prior art, the method has the advantages that the multi-modal problem is converted into the semi-supervised problem, and the mean-teach semi-supervised model is used for training the model, so that the prediction precision of the vehicle track can be effectively improved, and the problem of data imbalance in vehicle track prediction can be solved; the modal distribution in the vehicle prediction result can be effectively improved, and the trajectory prediction quality is remarkably improved.
本发明涉及一种基于半监督模型的车辆多模态轨迹预测方法,包括:基于道路地图信息获取车辆的未来可能意图和可能行驶的道路及中心线,并对目标车辆周围的道路地图信息以及目标车辆周围的其他车辆历史轨迹进行编码,随后使用半监督模型,为未来可能的目标车辆轨迹模态生成伪标签,并使用伪标签对不同模态下的预测轨迹进行监督学习,实现车辆轨迹的多模态预测。与现有技术相比,本发明将多模态问题转化为半监督问题,并使用mean‑teacher半监督模型对模型进行训练,能够有效得提升车辆轨迹的预测精度,并能解决车辆轨迹预测中存在的数据不平衡问题,能够有效改善车辆预测结果中的模态分布、显著提升轨迹预测质量。
Vehicle multi-modal trajectory prediction method based on semi-supervised model
一种基于半监督模型的车辆多模态轨迹预测方法
2023-06-27
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 / G01C Messen von Entfernungen, Höhen, Neigungen oder Richtungen , MEASURING DISTANCES, LEVELS OR BEARINGS / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS |
Multi-modal vehicle trajectory prediction method based on hierarchical order network
European Patent Office | 2023
|Interactive vehicle multi-modal trajectory prediction method based on GRU-GCN
European Patent Office | 2023
|