The invention discloses an intelligent driving behavior decision-making method and equipment fusing a complex network theory and a partially observable Markov decision-making process, provides an ideal decision-making model for multi-agent interaction in an uncertain environment by introducing the partially observable Markov decision-making process (POMDP), and provides an ideal decision-making model for multi-agent interaction along with the increase of model dimensions. In order to solve the problem of calculation complexity, a complex network cognition theory is introduced to model a driving environment for evaluating important nodes and extracting a risk tree, so that the confidence space of a POMDP model is reduced, model solving is accelerated, and the reliability of the driving environment is improved. Compared with a behavior decision-making method adopting a finite-state machine and an RL, the real-time behavior decision-making method has the advantages that the algorithm interpretability is ensured, the method can be effectively operated in the high-interaction dynamic uncertainty driving environment, the real-time behavior decision-making is realized, and the real-time behavior decision-making is realized in the high-interaction dynamic uncertainty driving environment. And a new solution thought is provided for development of a high-level intelligent driving system.

    本发明公开了融合复杂网络理论和部分可观察马尔可夫决策过程的智能驾驶行为决策方法及设备,通过引入部分可观察马尔可夫决策过程(POMDP),提供了一种在不确定性环境下,多智能体交互的理想决策模型,随着模型维度的增加,其计算复杂度也大大增加,为解决计算复杂度问题,本发明引入基于复杂网络认知理论对驾驶环境进行建模,用于评估重要节点并提取风险树,以缩小POMDP模型的置信空间、加速模型求解,从而实现在高度交互的动态不确定性驾驶环境下进行实时性的行为决策,相比于采用有限状态机和RL的行为决策方法,本发明在保证算法可解释性的同时,可在高度交互的动态不确定性驾驶环境中有效运行,为高级别的智能驾驶系统开发提供了新的解决思路。


    Access

    Download


    Export, share and cite



    Title :

    Intelligent driving behavior decision-making method and device fusing complex network theory and partially observable Markov decision-making process


    Additional title:

    融合复杂网络理论和部分可观察马尔可夫决策过程的智能驾驶行为决策方法及设备


    Contributors:
    CAI YINGFENG (author) / FU XINKE (author) / TENG CHENGLONG (author) / CHEN LONG (author) / WANG HAI (author) / LIU QINGCHAO (author) / SUN XIAOQIANG (author)

    Publication date :

    2023-04-28


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    G05D SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES , Systeme zum Steuern oder Regeln nichtelektrischer veränderlicher Größen / 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





    Position validation strategies using partially observable Markov decision processes

    Kochenderfer, Mykel J. / Shih, Kevin J. / Chryssanthacopoulos, James P. et al. | IEEE | 2011

    Free access

    State Estimation for Mobile Robot using Partially Observable Markov Decision Process

    Kang, D. / Hashimoto, H. / Harashima, F. | British Library Online Contents | 1996


    Signal Source Localization Using Partially Observable Markov Decision Processes

    Dressel, Louis K. / Kochenderfer, Mykel J. | AIAA | 2015