Liu, J., & Bihl, T. J., & Nagura, D. M. (2025, March), Reinforcement Learning with Human Experience (RLHE) for Racing Car Games  Paper presented at 2025 ASEE North Central Section  (NCS) Annual Conference, Marshall University, Huntington, West Virginia. https://peer.asee.org/54684

\bibitem{asee_peer_54684}
  Liu, J., \& Bihl, T. J., \& Nagura, D. M. (2025, March), \emph{Reinforcement Learning with Human Experience (RLHE) for Racing Car Games}
  Paper presented at 2025 ASEE North Central Section  (NCS) Annual Conference, Marshall University, Huntington, West Virginia.
  https://peer.asee.org/54684

Jundong Liu, Trevor Joseph Bihl, and Daniel Masami Nagura.  "Reinforcement Learning with Human Experience (RLHE) for Racing Car Games".  2025 ASEE North Central Section  (NCS) Annual Conference, Marshall University, Huntington, West Virginia, 2025, March.  ASEE Conferences, 2025.  https://peer.asee.org/54684 Internet.  25 Apr, 2025

\bibitem{asee_peer_54684}
  Jundong Liu, Trevor Joseph Bihl, and Daniel Masami Nagura.
  "Reinforcement Learning with Human Experience (RLHE) for Racing Car Games".
  \emph{2025 ASEE North Central Section  (NCS) Annual Conference, Marshall University, Huntington, West Virginia, 2025, March}.
  ASEE Conferences, 2025.
  https://peer.asee.org/54684 Internet.  25 Apr, 2025

@INPROCEEDINGS{asee_peer_54684
author = "Jundong Liu, Trevor Joseph Bihl, and Daniel Masami Nagura"
title = "Reinforcement Learning with Human Experience (RLHE) for Racing Car Games"
booktitle = "2025 ASEE North Central Section  (NCS) Annual Conference"
year = "2025"
month = "March"
address = "Marshall University, Huntington, West Virginia"
publisher = "ASEE Conferences"
note = {https://peer.asee.org/54684}
}

TY  - CPAPER
AB  - Path and motion planning are critical components of many autonomous tasks. In this work, we introduce a novel method to train reinforcement learning (RL) models for OpenAI Gym’s Car Racing game. The underlying playing policy is trained with a reinforcement learning (RL) setup, which consists states, actions, rewards and RL learning algorithms. Human players in- teract with the environment, generating an experience buffer as they play. During the training phase of the RL agents, the human experience buffer is combined with agent-generated data to establish a reinforcement learning with human experience (RLHE) paradigm. Experimental results show that our RLHE approach significantly accelerates the training process and en- hances the performance of RL agents, demonstrating marked improvements in race car motion planning.

AU  - Jundong Liu
AU  - Trevor Joseph Bihl
AU  - Daniel Masami Nagura
CY  - Marshall University, Huntington, West Virginia
DA  - 2025/03/28
PB  - ASEE Conferences
TI  - Reinforcement Learning with Human Experience (RLHE) for Racing Car Games
UR  - https://peer.asee.org/54684
ER  -