The invention provides an automatic driving vehicle behavior prediction system and method fusing a complex network and a graph Transformer, and the method comprises the steps: firstly constructing a vehicle interaction complex network, and taking original map information data, vehicle motion data and local interaction information as the multi-layer input of a model; then, by means of graph convolution of a graph Transform and a multi-node attention mechanism, modeling and learning are conducted on the complex relation between the vehicles, and global and local correlation of the vehicles is obtained; and finally, through training the prediction model, enabling the graph Transform model to have good generalization and robustness, and accurately outputting the behavior of the automatic driving vehicle.
本发明提供了一种融合复杂网络和图Transformer的自动驾驶车辆行为预测系统及方法,首先,构建车辆交互复杂网络,同时将原始地图信息数据、车辆的运动数据、局部交互信息作为模型的多层输入;然后,借助图Transformer的图卷积和多头节点注意力机制,对车辆之间的复杂关系进行建模和学习,获取车辆全局、局部的相关性;最后,通过训练预测模型,使得图Transformer模型具有良好的泛化性和鲁棒性,可以准确输出自动驾驶车辆的行为。
Automatic driving vehicle behavior prediction method and system fusing complex network and graph Transformer
一种融合复杂网络和图Transformer的自动驾驶车辆行为预测方法及系统
2023-10-10
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 / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen |
European Patent Office | 2022
|Distilled Routing Transformer for Driving Behavior Prediction
SAE Technical Papers | 2023
|Freight train automatic driving control system and method fusing adjacent train information
European Patent Office | 2024
|European Patent Office | 2023
|