Investigating Successor Features in the Domain of Autonomous Vehicle Control

In this article, the basic Reinforcement Learning (RL) concepts are discussed, continued with a brief explanation of Markov Decision Processes (MDPs). Reasoning for the application of RL in the autonomous vehicle control domain is accompanied with a developed basic environment for simulation-based training of agents. Furthermore, the authors look at the available literature of successor features, and the recent achievements of its utilization. The authors motivation is to tackle the problem of credit assignment with reward decomposition by using successor features, because the complex tasks while driving can cause unsuccessful training and can be challenging. After explaining the applied methodology and showing how it works, state-of-the-art ideas are investigated and infused into the vehicle control realm. Moreover, the paper details how these features can be tailored to the highway driving scenarios and what is the secret behind its capability to boost the performance of the RL agent. In order to investigate the proposed problems in a credible way the authors applied a high-fidelity traffic simulator (SUMO) as the authors environment, and concluded different trainings based on various scenarios. The authors present successor features applied to an autonomous vehicle control setting, such as highway commute. The authors results imply that learned skills can help with the multi-objective rewarding problem, and agents applied to changing reward systems can adapt quickly to the new tasks. The only thing to find is the correct decomposition and selection of successor features.

• Record URL:
  • https://doi.org/10.1016/j.trpro.2022.02.023
• Record URL:
  • http://www.sciencedirect.com/science/article/pii/S2352146522001508
• Availability:
  • Find a library where document is available. Order URL: http://worldcat.org/issn/23521465
• Supplemental Notes:
  • © 2022 Laszlo Szoke, et al. Published by Elsevier B.V. Abstract reprinted with permission of Elsevier.
• Authors:
  • Szoke, Laszlo
  • Aradi, Szilard
  • Tettamanti, Tamas
• Conference:
  • 24th EURO Working Group on Transportation Meeting, EWGT 2021
  • Location: Aveiro , Portugal
  • Date: 2021-9-8 to 2021-9-10
• Publication Date: 2022

Language

• English

Media Info

• Media Type: Web
• Features: References;
• Pagination: pp 181-188
• Serial:
  • Transportation Research Procedia
  • Volume: 62
  • Issue Number: 0
  • Publisher: Elsevier
  • ISSN: 2352-1465
  • Serial URL: http://www.sciencedirect.com/science/journal/23521465/
• Publication flags:

  Open Access (libre)

Subject/Index Terms

• TRT Terms: Advanced vehicle control systems; Autonomous vehicles; Forecasting; Learning; Markov processes; Traffic simulation
• Identifier Terms: SUMO (Traffic simulation model)
• Subject Areas: Data and Information Technology; Highways; Operations and Traffic Management; Vehicles and Equipment;

Filing Info

• Accession Number: 01842365
• Record Type: Publication
• Files: TRIS
• Created Date: Apr 13 2022 9:37AM