A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Intelligent Decision Framework of Shield Attitude Correction Based on Deep Reinforcement Learning
Shield attitude is an important control element in the construction of shield projects. Due to the complex shield construction environment and the difficulties of shield machine operation, shield attitude control often has lag and inaccuracy. For the problem of shield snaking motion caused by untimely and inaccurate shield attitude adjustment, scholars have proposed some attitude prediction models and fuzzy PID control models based on machine learning and deep learning. However, these models are highly environment-dependent and require expert priori knowledge, making it difficult to predict or control the shield attitude in time when the construction environment changes. To address the shortcomings of these models, this paper proposes an intelligent decision framework of shield attitude correction (IDFSAC) based on deep reinforcement learning. This framework is able to interact with the construction environment and adaptively find the best correction and control strategy. The effectiveness of IDFSAC is experimented using a shield construction simulation system. Results reveal that the framework can adapt to the changes of construction environment and realize automatic and intelligent correction and control of shield attitude.
Intelligent Decision Framework of Shield Attitude Correction Based on Deep Reinforcement Learning
Shield attitude is an important control element in the construction of shield projects. Due to the complex shield construction environment and the difficulties of shield machine operation, shield attitude control often has lag and inaccuracy. For the problem of shield snaking motion caused by untimely and inaccurate shield attitude adjustment, scholars have proposed some attitude prediction models and fuzzy PID control models based on machine learning and deep learning. However, these models are highly environment-dependent and require expert priori knowledge, making it difficult to predict or control the shield attitude in time when the construction environment changes. To address the shortcomings of these models, this paper proposes an intelligent decision framework of shield attitude correction (IDFSAC) based on deep reinforcement learning. This framework is able to interact with the construction environment and adaptively find the best correction and control strategy. The effectiveness of IDFSAC is experimented using a shield construction simulation system. Results reveal that the framework can adapt to the changes of construction environment and realize automatic and intelligent correction and control of shield attitude.
Intelligent Decision Framework of Shield Attitude Correction Based on Deep Reinforcement Learning
Lecture Notes in Civil Engineering
Geng, Guoqing (editor) / Qian, Xudong (editor) / Poh, Leong Hien (editor) / Pang, Sze Dai (editor) / Xu, J. (author) / Bu, J. F. (author) / Zhang, L. G. (author) / Zhang, J. (author) / Li, K. F. (author) / Liu, S. M. (author)
Proceedings of The 17th East Asian-Pacific Conference on Structural Engineering and Construction, 2022 ; Chapter: 102 ; 1273-1287
2023-03-14
15 pages
Article/Chapter (Book)
Electronic Resource
English
Shield construction management , Shield attitude correction , Intelligent decision , Reinforcement learning , Deep learning Engineering , Building Construction and Design , Structural Materials , Solid Mechanics , Sustainable Architecture/Green Buildings , Light Construction, Steel Construction, Timber Construction , Offshore Engineering
Dynamic coordinated control of attitude correction for the shield tunneling based on load observer
| Online Contents | 2012
Dynamic coordinated control of attitude correction for the shield tunneling based on load observer
| British Library Online Contents | 2012