The invention is suitable for the technical field of intelligent driving, and provides an end-to-end automatic driving lane changing decision-making method based on multi-modal input, and the method comprises the following steps: state information collection: collecting an aerial view, a semantic image, and state information of a vehicle and surrounding vehicles, and carrying out the data processing; constructing a multi-modal input network; determining the architecture of the deep neural network; a convolutional neural network is adopted to process image data, and a full connection layer is adopted to process state information; designing a state space, an action space and a reward function; and end-to-end training: carrying out end-to-end training on the constructed multi-modal input network by using the DDPG. According to the invention, data from different sensors are utilized at the same time, so that more comprehensive and accurate environment perception and lane changing decision are realized; different modal input data are effectively fused, and an end-to-end deep learning network is adopted for learning and decision making, so that information loss and coupling among a plurality of modules in a traditional method are avoided, and the stability and reliability of the system are improved.
本发明适用于智能驾驶技术领域,提供了一种基于多模态输入的端到端自动驾驶换道决策方法,包括以下步骤:状态信息收集:收集鸟瞰图、语义图像以及自车和周车的状态信息,进行数据处理;构建多模态输入网络:确定深度神经网络的架构;采用卷积神经网络来处理图像数据,采用全连接层来处理状态信息;设计状态空间、动作空间、奖励函数;端到端训练:使用DDPG对构建的多模态输入网络进行端到端的训练。本发明同时利用来自不同传感器的数据,实现了更全面、准确的环境感知和换道决策;将不同模态输入数据进行有效融合,并采用端到端的深度学习网络进行学习和决策,避免了传统方法中多个模块之间的信息丢失和耦合,提高了系统的稳定性和可靠性。
End-to-end automatic driving lane changing decision-making method based on multi-modal input
一种基于多模态输入的端到端自动驾驶换道决策方法
2024-10-29
Patent
Electronic Resource
Chinese
IPC: | G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / 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 / G01S RADIO DIRECTION-FINDING , Funkpeilung / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G06V |
Automatic driving vehicle lane changing decision-making method based on multi-index evaluation
European Patent Office | 2024
|Automatic driving lane changing decision making method and device and vehicle
European Patent Office | 2021
|Automatic driving intelligent acceleration and deceleration lane changing decision-making method
European Patent Office | 2025
|Differential automatic lane changing decision-making method
European Patent Office | 2023
|Hybrid automatic driving lane changing decision-making method based on deep reinforcement learning
European Patent Office | 2024
|