A responsive traffic signal control method and an apparatus therefor are disclosed. The responsive traffic signal control method comprises the steps of: (a) inputting a plurality of intersection images including a current traffic state into a deep convolutional neural network (CNN) model so as to respectively output Q-function values that are an approximation of cumulative expected corrections according to changes in the traffic state for each of the current traffic state and respective signal control operation pairs; (b) selecting a signal control operation corresponding to the maximum value of the derived Q-function values as a target signal control operation; (c) executing the selected target signal control operation; (d) updating the correction value using changes in the number of vehicles waiting at and approaching a intersection according to the execution of the target signal control operation; and (e) updating parameter values of the Q-function using the correction value and the Q-function values.

    L'invention concerne un procédé de commande de signal de trafic de réponse et un appareil associé. Les étapes de procédé de commande de signal de trafic de réponse consistent : (a) à entrer une pluralité d'images d'intersection incluant un état actuel de trafic dans un modèle de réseau neural convolutif (CNN) profond de manière à transmettre respectivement des valeurs de fonction Q qui sont une approximation de corrections prévues cumulatives selon des changements de l'état de trafic pour chacune des paires d'état actuel de trafic et d'opérations de commande de signal respectif; (b) à sélectionner une opération de commande de signal correspondant à la valeur maximale des valeurs de fonction Q dérivées en tant qu'opération cible de commande de signal; (c) à exécuter l'opération cible de commande de signal sélectionnée; (d) à actualiser la valeur de correction utilisant des changements du nombre de véhicules attendant à une intersection et s'en approchant selon l'exécution de l'opération cible de commande de signal; et (e) à actualiser des valeurs de paramètres de la fonction Q utilisant la valeur de correction et les valeurs de fonction Q.

    감응식 교통 신호 제어 방법 및 그 장치가 개시된다. 감응식 교통 신호 제어 방법은, (a) 현재 교통 상태를 포함하는 복수의 교차로 이미지를 딥 CNN(deep CNN) 모델에 입력함으로써, 상기 현재 교통 상태와 각 신호 제어 동작 페어(pair) 각각에 대한 교통 상태 변화에 따른 누적기대보상의 근사화된 평가 함수값(Q-function)을 각각 출력하는 단계;(b) 상기 도출된 평가 함수값들 중 최대값에 상응하는 신호 제어 동작을 타겟 신호 제어 동작으로 선택하는 단계; (c) 상기 선택된 타겟 신호 제어 동작을 실행하는 단계; (d) 상기 타겟 신호 제어 동작 실행에 따른 교차로 대기 및 접근 차량수 변화를 이용하여 보상값을 갱신하는 단계; 및 (e) 상기 보상값과 상기 평가 함수값을 이용하여 상기 평가 함수의 파라미터값을 갱신하는 단계를 포함한다.


    Access

    Download


    Export, share and cite



    Title :

    RESPONSIVE TRAFFIC SIGNAL CONTROL METHOD AND APPARATUS THEREFOR


    Additional title:

    PROCÉDÉ DE COMMANDE DE SIGNAL DE TRAFIC DE RÉPONSE ET APPAREIL ASSOCIÉ
    감응식 교통 신호 제어 방법 및 그 장치


    Contributors:

    Publication date :

    2018-02-15


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Korean


    Classification :

    IPC:    G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS



    Demand-responsive traffic signal control research

    Gartner, Nathan H. | Elsevier | 1985


    Person-Based Traffic Responsive Signal Control Optimization

    Christofa, Eleni / Papamichail, Ioannis / Skabardonis, Alexander | IEEE | 2013




    Responsive Signal Control Design for Urban Traffic Network

    Li, Ying / Chow, Andy H. F. | ASCE | 2012