Enhanced Image Preprocessing Method for an Autonomous Vehicle Agent System

Kaisi Huang1, Mingyun Wen2, Jisun Park3, Yunsick Sung4, Jong Hyuk Park5 and Kyungeun Cho6

  1. Department of Multimedia Engineering, Dongguk University-Seoul
    30, Pildongro-1-gil, Jung-gu, Seoul 04620, Republic of Korea
    hkshks0825@dongguk.edu
  2. Department of Multimedia Engineering, Dongguk University-Seoul
    30, Pildongro-1-gil, Jung-gu, Seoul 04620, Republic of Korea
    wmy_ncut@dongguk.edu
  3. Department of Multimedia Engineering, Dongguk University-Seoul
    30, Pildongro-1-gil, Jung-gu, Seoul 04620, Republic of Korea
    jisun@dongguk.edu
  4. Department of Multimedia Engineering, Dongguk University-Seoul
    30, Pildongro-1-gil, Jung-gu, Seoul 04620, Republic of Korea
    sung@mme.dongguk.edu
  5. Department of Computer Science and Engineering, Seoul National University of Science and Technology (SeoulTech)
    232 Gongneung-ro, Nowon-gu, Seoul, 01811, Korea
    jhpark1@seoultech.ac.kr
  6. Department of Multimedia Engineering, Dongguk University-Seoul
    30, Pildongro-1-gil, Jung-gu, Seoul 04620, Republic of Korea
    cke@dongguk.edu

Abstract

Excessive training time is a major issue face when training autonomous vehicle agents with neural networks by using images as input. This paper proposes a deep time-economical Q network (DQN) input image preprocessing method to train an autonomous vehicle agent in a virtual environment. The environmental information is extracted from the virtual environment. A top-view image of the entire environment is then redrawn according to the environmental information. During training of the DQN model, the top-view image is cropped to place the vehicle agent at the center of the cropped image. The current frame top-view image is combined with the images from the previous two training iterations. The DQN model use this combined image as input. The experimental results indicate higher performance and shorter training time for the DQN model trained with the preprocessed images compared with that trained without preprocessing.

Key words

Image preprocessing, Reinforcement learning, Deep Q learning

Digital Object Identifier (DOI)

https://doi.org/10.2298/CSIS200212005H

Publication information

Volume 18, Issue 2 (April 2021)
Special Issue on Emerging Services in the Next-Generation Web: Human Meets Artificial Intelligence
Year of Publication: 2021
ISSN: 2406-1018 (Online)
Publisher: ComSIS Consortium

Full text

DownloadAvailable in PDF
Portable Document Format

How to cite

Huang, K., Wen, M., Park, J., Sung, Y., Park, J. H., Cho, K.: Enhanced Image Preprocessing Method for an Autonomous Vehicle Agent System. Computer Science and Information Systems, Vol. 18, No. 2, 461–479. (2021), https://doi.org/10.2298/CSIS200212005H