Deep Reinforcement Learning based Automated UAV Path Planning Model for Worker Safety Monitoring in Construction

Dohyeong Kim; Yuna Jung; Dongmin Lee

doi:10.55785/JCAR.2.4.17

All Issue

2023 Vol.2, Issue 4 Preview Page Next Page

Research Article

Deep Reinforcement Learning based Automated UAV Path Planning Model for Worker Safety Monitoring in Construction 건설현장 근로자 안전 모니터링을 위한 심층강화학습 기반 UAV 경로탐색 자동화 모델 개발

29 December 2023. pp. 17-24

PDF XML

Abstract

Unmanned aerial vehicles (UAVs) have emerged as an effective solution to address the limitations of labor-intensive safety monitoring at construction sites. However, UAV-based monitoring methods exhibit reduced path-planning efficiency in dynamic construction environments. To overcome this limitation, this study develops an automated UAV path-planning model based on deep reinforcement learning (DRL). The model autonomously learns to navigate towards randomly moving workers while avoiding obstacles in a simulated construction space, employing a reward-punishment system. A case study compares the effectiveness of the proposed model in ever-changing dynamic construction environments with that of a GPS navigation algorithm, similar to existing UAV path-planning approaches. The results demonstrate that the model can identify flight paths to monitor workers more quickly than the GPS navigation algorithm in dynamic environments where workers move randomly. This suggests that DRL-based automated UAV path-planning holds significant potential for the application of UAVs in worker safety monitoring at dynamic construction sites.

Keywords

Construction Safety Monitoring

Unmanned Aerial Vehicle

Deep Reinforcement Learning

Path Planning

최근 UAV 기반의 건설현장 안전 모니터링 방법이 기존 인력 기반 모니터링의 한계점을 해결하는 유용한 대안으로 부상하고 있다. 그러나 UAV 기반 건설현장 안전 모니터링 방법은 동적인 건설 환경에서 경로탐색 효율이 저하한다. 이러한 한계점을 해결하기 위해, 본 연구는 심층강화학습 기반 UAV 경로탐색 자동화 모델을 개발한다. 제안한 모델은 가상의 건설 현장 공간에서 장애물을 회피하며 무작위로 움직이는 근로자를 향한 경로 생성을 자동화하는 방법을 강화학습 내의 보상과 처벌 시스템을 통해 스스로 학습한다. 저자들은 사례 연구를 통해 경로탐색 자동화 모델이 기존 UAV의 경로탐색과 유사한 GPS 내비게이션 알고리즘에 비해 지속적으로 변화하는 동적인 건설 환경 상황에서 더 효과적인 경로탐색을 하는지 비교하였다. 비교 결과, 제안한 모델은 여러명의 근로자가 무작위적으로 움직이는 동적 환경에서 GPS 내비게이션 알고리즘보다 더 빠르게 근로자를 탐색할 수 있는 비행경로를 탐색하고, 모니터링 할 수 있었다. 이는 동적인 건설현장에서 UAV를 근로자 안전 모니터링에 적용함에 있어 심층강화학습 기반의 경로탐색 자동화 모델이 뛰어난 잠재력을 가지고 있음을 시사한다.

키워드

건설 안전 모니터링

무인 비행 장치

심층강화학습

경로탐색

References

Alizadehsalehi, S., Yitmen, I., Celik, T., and Arditi, D. (2018). The effectiveness of an integrated BIM/UAV model in managing safety on construction sites. International Journal of Occupational Safety and Ergonomics, 26, pp. 829-844. 10.1080/10803548.2018.150448730043680

Fang, W., Ding, L., and Luo, P. E. D. H. (2018). Falls from heights: A computer vision-based approach for safety harness detection. Automation in Construction, 91, pp. 53-61. 10.1016/j.autcon.2018.02.018

Irizarry, J., Gheisari, M., and Walker, B. N. (2012). Usability assessment of drone technology as safety inspection tools. Journal of Information Technology in Construction. 17(12), pp. 194-212.

Jung, W. (2020). Build reinforcement learning AI process for cooperative play with users. Journal of Korea Game Society, 20(1), pp. 57-66. 10.7583/JKGS.2020.20.1.57

Khan, M., Khalid, R., Anjum, S., Khan, N., and Park, C. (2021). IMU based smart safety hook for fall prevention at construction sites. 2021 IEEE Region 10 Symposium, Jeju, Korea, pp. 1-6. 10.1109/TENSYMP52854.2021.9550944

Khan, N., Zaidi, S. F. A., Yang, J., Park, C., and Lee, D. (2023). Construction work-stage-based rule compliance monitoring framework using computer vision (CV) technology. Buildings, 13(8), 2093 p. 10.3390/buildings13082093

Kim, D., Liu, M., Lee, S., and Kamat, V. R. (2019). Remote proximity monitoring between mobile construction resources using camera- mounted UAVs. Automation in Construction, 99, pp. 168-182. 10.1016/j.autcon.2018.12.014

Korea Safety Education Association. (2022). Let's learn about safety rules at construction sites. [Korea Safety Education Association]. Naver Blog, https://post.naver.com/viewer/postView.naver?volumeNo=34163402&memberNo=19462446 (Accessed November 16, 2023)

Lee, J. H., Park, J. H., and Jang, B. T. (2018). Design of robot based work progress monitoring system for the building construction site. International Conference on Information and Communication Technology Convergence (ICTC), Jeju, Korea, pp. 1420-1422. 10.1109/ICTC.2018.8539444

Levitt, R., and Samelson, N. (1993). Construction safety management (Second edition). John Wiley & Sons, New York. pp. 23-44.

Ministry of Employment and Labor of South Korea. (2022). Announcement of the status of industrial accident deaths in 2021. Press release (March 15). https://www.moel.go.kr/common/downloadFile.do?file_seq=20220301130&bbs_seq=13330&bbs_id=12 (Accessed November 11, 2023).

Pittokopiti, M., and Grammenos, R. (2019). Infrastructureless UWB based collision avoidance system for the safety of construction workers. 2019 26th International Conference on Telecommunications (ICT), Hanoi, Vietnam, pp. 490-495. 10.1109/ICT.2019.8798845

Ryu, U., and Kim, Y. (2022). A study on the dual control platform for drone field training. Journal of Platform Technology, 10(2), pp. 20-26.

Shim, W. I., Park, T. H., and Kim, K. J. (2018). Comparison of policy optimization reinforcement learning for simulated autonomous car environment. Proceedings of the Korean Society of Information Sciences, pp. 833-835.

Yu, L., Huang, M. M., Jiang, S., Wang, C., and Wu, M. (2023). Unmanned aircraft path planning for construction safety inspections. Automation in Construction, 154(105005), pp. 1-19. 10.1016/j.autcon.2023.105005

Information

Publisher :Korean Society of Automation and Robotics in Construction
Publisher(Ko) :(사)한국건설자동화·로보틱스학회
Journal Title :Journal of Construction Automation and Robotics
Journal Title(Ko) :건설자동화·로보틱스 논문집
Volume : 2
No :4
Pages :17-24
Received Date : 2023-11-23
Revised Date : 2023-12-14
Accepted Date : 2023-12-18
DOI :https://doi.org/10.55785/JCAR.2.4.17

Journal of Construction Automation and Robotics ISSN:2800-0552(Print) 2951-116X(Online) 건설자동화·로보틱스 논문집

All Issue