To solve the problem in which, although autonomous driving is dealt with reinforcement learning algorithms based on the Q-learning or policy gradient, rule-based methods cannot change scale well whereas reinforcement learning methods based on the Q-learning or policy gradient are not sample-efficient and are very time consuming to solve because of the need to explore a large state space.SOLUTION: A framework combines vision and sample-efficient reinforcement-learning on the basis of guided policy search for autonomous driving. A controller extracts environmental information from vision and is trained to drive using reinforcement learning.SELECTED DRAWING: Figure 1

    【課題】Q学習又は方策勾配法に基づいた強化学習アルゴリズムを用いて全自動運転に取り組んでいる。しかしながら、規則に基づいた方法はスケール変更がうまくできず、その一方、Q学習または方策勾配法に基づいた強化学習法は、大きな状態空間を調査する必要があるせいで、サンプル効率的でなく解決するために非常に時間がかかる。【解決手段】フレームワークは、全自動運転のガイド付き方策検索に基づいて、視界とサンプル効率的な強化学習とを組み合わせる。制御装置は、視界から環境情報を抽出し、強化学習を用いて運転するように訓練される。【選択図】図1


    Access

    Download


    Export, share and cite



    Title :

    VISION-BASED SAMPLE-EFFICIENT REINFORCEMENT LEARNING FRAMEWORK FOR AUTONOMOUS DRIVING


    Additional title:

    全自動運転用の視界によるサンプル効率的な強化学習のフレームワーク


    Contributors:

    Publication date :

    2019-10-24


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Japanese


    Classification :

    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



    - - VISION-BASED SAMPLE-EFFICIENT REINFORCEMENT LEARNING FRAMEWORK FOR AUTONOMOUS DRIVING

    CHIANG SU HUI / LIU MING CHANG | European Patent Office | 2019

    Free access


    Autonomous Driving with Deep Reinforcement Learning

    Zhu, Yuhua / Technische Universität Dresden | SLUB | 2023


    DQN-Based Deep Reinforcement Learning for Autonomous Driving

    Pérez-Gil, Óscar / Barea, Rafael / López-Guillén, Elena et al. | Springer Verlag | 2020


    Efficient-Enhanced Reinforcement Learning for Autonomous Driving in Urban Traffic Scenarios

    Yin, Jianwen / Jiang, Zhengmin / Liang, Qingyi et al. | IEEE | 2023