SRFofSBDDQN

Implementation of Setting Reward Function of Sensor Based DDQN Model (SRFofSBDDQN) article.

Environment

Environment created with Pygame module. It has 5 discrete actions. It has 2 different rewarding. Main purpose of the environment is passing 100 objects/obstacles which are created with random size and positions.

Action Space

Ind	Act
0	Throttle
1	Brake
2	Left
3	Right
4	Nothing

Model Training and Testing

Model created in gNet which created in pure Python based on NumPy.

@research article { 
    ejosat1008702, 
    journal = {Avrupa Bilim ve Teknoloji Dergisi}, 
    issn = {}, 
    eissn = {2148-2683}, 
    address = {}, 
    publisher = {Osman SAĞDIÇ}, 
    year = {2021}, 
    volume = {}, 
    pages = {539 - 544}, 
    doi = {10.31590/ejosat.1008702}, 
    title = {Setting Reward Function of Sensor Based DDQN Model}, 
    key = {cite}, 
    author = {Kabataş, Mehmet Gökçay and İlhan Omurca, Sevinç} }