The invention provides an automatic driving vehicle decision making method based on driving experience self-learning. The method comprises the following two basic steps: S1, designing an upper behavior planning layer; an MDP is used to complete a vehicle automatic driving decision, that is, a Markov decision process is used to complete vehicle behavior planning; s2, designing a lower motion planning control layer; a non-linear model predictive control technology is introduced to apply rigid constraint to a target automatic driving vehicle, and an upper behavior planning layer autonomously learns a vehicle control instruction based on observation information collected from a driving environment and then transmits the vehicle control instruction to model predictive control, namely, the vehicle control instruction is transmitted to an MPC; the MPC addresses the tracking optimization problem and provides low level commands to the target autonomous vehicle. According to the invention, integration of upper-layer behavior planning and lower-layer motion planning is realized, the training efficiency of vehicle automatic driving is greatly improved, and the performance and applicability of vehicle automatic driving in a real-world driving environment are improved.
本发明提供了一种基于驾驶经验自学习的自动驾驶车辆决策制定方法,两个基本步骤:S1,上层行为规划层设计;使用MDP完成车辆自动驾驶决策,亦即使用马尔可夫决策过程完成车辆行为规划;S2,下层运动规划控制层设计;通过引入非线性模型预测控制技术对目标自动驾驶车辆施加刚性约束,上层行为规划层将基于从驾驶环境中收集的观测信息自主学习车辆控制指令,随后将车辆控制指令传递给模型预测控制,亦即将车辆控制指令传递给MPC,MPC解决跟踪优化问题并提供低等级命令给目标自动驾驶车辆。本发明实现了上层行为规划和下层运动规划的集成,大幅度提高了车辆自动驾驶的训练效率,提升了车辆自动驾驶在真实世界驾驶环境中的性能和适用性。
Automatic driving vehicle decision making method based on driving experience self-learning
基于驾驶经验自学习的自动驾驶车辆决策制定方法
2024-12-13
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 |
Automatic driving vehicle microscopic decision-making method based on reinforcement learning
European Patent Office | 2020
|European Patent Office | 2024
|Automatic driving decision-making method based on driving intensity
European Patent Office | 2025
|European Patent Office | 2023
|Automatic driving decision-making method and system, medium, automatic driving system and vehicle
European Patent Office | 2024
|