Simultaneous generation of assembly paths and sequences for shipbuilding piping using modified RRT with UCB strategy

Read full text
Authors: Tomoyuki Taniguchi, Kohei Matsuo
Cited by: 1
References: 26

Abstract

Abstract In this study, we propose a method for the automatic generation of assembly sequences and assembly paths for ship piping. Planning and verifying the assembly procedures of piping components in advance is a critical process, as it directly impacts cost, schedule, and quality. However, manual planning requires significant time and effort due to the large number of components and the necessity of considering interferences. To address this challenge, this paper proposes a novel method capable of simultaneously generating both assembly paths and assembly sequences. First, for piping in shipbuilding, we demonstrate that piping installation through through-holes constitutes a narrow-passage problem. This type of problem is challenging for conventional Rapidly-exploring Random Trees (RRT) methods. To address this issue, we propose a modified-RRT algorithm that improves sampling efficiency by incorporating collision resolution operations. Experimental results using actual piping data demonstrated that the proposed method achieved a success rate of over 95% in generating assembly paths, representing a significant improvement over conventional method. Second, in generating assembly procedures, some piping components cannot be installed later or are difficult to install due to interference; therefore, assembly paths cannot be calculated efficiently. Moreover, since the expected outcomes of RRT cannot be predicted in advance, maintaining a balance between exploration and exploitation is crucial. Therefore, we propose a method to efficiently generate both assembly sequences and assembly paths by utilizing intermediate results from RRT and selecting piping components based on the Upper Confidence Bound (UCB) strategy. A case study was conducted for an engine room compartment containing numerous piping components using actual ship model data. As a result, for problems involving hundreds of pipes that could not be solved within one hour using conventional methods, the proposed method was able to compute assembly paths for all pipes within approximately 10–15 min. Based on these results, we conclude that the modified-RRT and UCB-based assembly path and procedure generation methods can efficiently determine assembly procedures for shipbuilding piping.

Fields of Study & Subjects

Manufacturing Process and Optimization Optimization and Packing Problems Robot Manipulation and Learning

Publication details

Journal / source
Journal of Marine Science and Technology
Publisher
Springer Science and Business Media LLC
Volume
31
Issue
1
Pages
124–137
Publication date
19 January 2026
Language
EN