로봇학회논문지 (The Journal of Korea Robotics Society)

  • 제19권2호
  • /
  • Pages.196-202
  • /
  • 2024
  • /
  • 1975-6291(pISSN)
  • /
  • 2287-3961(eISSN)
한국로봇학회 (Korea Robotics Society)
권호 표지
DOI QR코드

DOI QR Code

그래프 기반 상태 표현을 활용한 작업 계획 알고리즘 개발

Task Planning Algorithm with Graph-based State Representation

  • Seongwan Byeon (Department of Electronic Engineering, Hanyang University) ;
  • Yoonseon Oh (Department of Electronic Engineering, Hanyang University)
  • 투고 : 2024.03.08
  • 심사 : 2024.04.23
  • 발행 : 2024.05.31
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

The ability to understand given environments and plan a sequence of actions leading to goal state is crucial for personal service robots. With recent advancements in deep learning, numerous studies have proposed methods for state representation in planning. However, previous works lack explicit information about relationships between objects when the state observation is converted to a single visual embedding containing all state information. In this paper, we introduce graph-based state representation that incorporates both object and relationship features. To leverage these advantages in addressing the task planning problem, we propose a Graph Neural Network (GNN)-based subgoal prediction model. This model can extract rich information about object and their interconnected relationships from given state graph. Moreover, a search-based algorithm is integrated with pre-trained subgoal prediction model and state transition module to explore diverse states and find proper sequence of subgoals. The proposed method is trained with synthetic task dataset collected in simulation environment, demonstrating a higher success rate with fewer additional searches compared to baseline methods.

키워드

과제정보

This work was supported by Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government (MSIT) (No. 2022-0-00907, Development of AI Bots Collaboration Platform and Self-organizing AI, 80% and No.RS-2020-II201373, Artificial Intelligence Graduate School Program (Hanyang University), 20%).

참고문헌

  1. M. D. Graaf, S. B. Allouch, and J. V. Diik, "Why do they refuse to use my robot?: Reasons for non-use derived from a long-term home study," 2017 ACM/IEEE International Conference on Human-Robot Interaction, Vienna, Austria, pp. 224-233, 2017, DOI: 10.1145/2909824.3020236.
  2. H. Guo, F. Wu, Y. Qin, R. Li, K. Li, and K. Li, "Recent Trends in Task and Motion Planning for Robotics: A Survey," ACM Comput. Surv, vol. 55, no. 13, pp. 1-36, Jul., 2023, DOI: 10.1145/3583136.
  3. M. Ghallab, C. A. Knoblock, D. E. Wilkins, A. Barrett, D. Christianson, M. T. Friedman, C. Kwok, K. Golden, S. Penberthy, D. E. Smith, Y. Sun, and D. Weld, "PDDL - The Planning Domain Definition Language," Technical report, Yale Center for Computational Vision and Control, 1998, [Online], https://www.researchgate.net/publication/2278933.
  4. C. Paxton, Y. Barnoy, K. Katyal, R. Arora, and G. D. Hager, "Visual robot task planning," 2019 International conference on robotics and automation (ICRA), pp. 8832-8838, 2019, DOI: 10.1109/ICRA.2019.8793736.
  5. T. Silver, V. Hariprasad, R. S. Shuttleworth, N. Kumar, T. Lozano-Perez, and L. P. Kaelbling, "PDDL planning with pretrained large language models," NeurIPS 2022 foundation models for decision making workshop, 2022, [Online], https://openreview.net/forum?id=1QMMUB4zfl, Accessed: Apr. 23, 2024.
  6. J. Zhou, G. Cui, S. Hu, Z. Zhang, C. Yang, Z. Liu, L. Wang, C. Li, and M. Sun, "Graph neural networks: A review of methods and applications," AI open, vol. 1, pp.57-81, 2020, DOI: 10.1016/j.aiopen.2021.01.001.
  7. T. Silver, R. Chitnis, A. Curtis, J. Tenenbaum, T. Lozano-Perez, and L. P. Kaelbling, "Planning with learned object importance in large problem instances using graph neural networks," Proceedings of the AAAI conference on artificial intelligence, pp. 11962-11971, Sept., 2021, DOI: 10.48550/arXiv.2009.05613.
  8. K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 770-778, 2016, DOI: 10.48550/arXiv.1512.03385.
  9. J. J. Kuffner and S. M. LaValle, "RRT-connect: An efficient approach to single-query path planning," Proceedings 2000 ICRA. Millennium Conference. IEEE International Conference on Robotics and Automation. Symposia Proceedings (Cat. No. 00CH37065), San Francisco, CA, USA, pp. 995-1001, 2000, vol. 2, DOI: 10.1109/ROBOT.2000.844730.