The invention belongs to the field of automatic driving, and particularly provides an end-to-end automatic driving decision-making method based on driving priori knowledge, which is used for solving the problems of insufficient information, low interpretability of an end-to-end model and the like caused by single sensor input in the prior art. According to the method, lane line features are extracted as priori knowledge, then the lane line priori knowledge and an original driving image are input together, a time sequence is constructed by combining current input and previous moment input, and finally an automatic driving decision-making corner at the current moment is output through an end-to-end decision-making network based on the time sequence. In the process, the lane line features serve as priori knowledge and participate in decision making together with the original driving image, and meanwhile, current input and previous input participate in decision making together, so that decision making precision is remarkably improved, the precision and safety of an automatic driving system are improved, and research, development, popularization and use of an automatic driving vehicle are facilitated.
本发明属于自动驾驶领域,具体提供一种基于驾驶先验知识的端到端自动驾驶决策方法,用以解决现有技术中由于单传感器输入导致的信息不足、端到端模型的可解释性低等问题。本发明将车道线特征作为先验知识进行提取,再将车道线先验知识与原始驾驶图像进行共同输入,并结合时序构造将当前输入与之前时刻的输入构造的时间序列,最终基于时间序列由端到端决策网络输出当前时刻的自动驾驶决策转角;在该过程中,车道线特征特征作为先验知识与原始驾驶图像共同参与决策,同时将当前输入与之前时刻的输入共同参与决策,使得决策精度显著提升,进而提高自动驾驶系统的精确度与安全性,有助于自动驾驶车辆的研发、推广及使用。
End-to-end automatic driving decision-making method based on driving priori knowledge
一种基于驾驶先验知识的端到端自动驾驶决策方法
2023-04-14
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 / G06V |
Automatic driving decision-making method based on driving intensity
European Patent Office | 2025
|Inplausible automatic driving decision-making method based on causal knowledge
European Patent Office | 2025
|Automatic driving vehicle decision making method based on driving experience self-learning
European Patent Office | 2024
|European Patent Office | 2020
|European Patent Office | 2024
|