Objective Shield method is a common method for subway tunnel excavation. In shield construction, the segment assembly process is mostly completed by manual operation. In the case of complex motion conditions of segment erector, manual operation of the erector seriously affects the efficiency and quality of segment assembly, and brings hidden dangers to safety construction. Therefore, it is necessary to realize the automatic assembly of shield segments, and improve the efficiency and accuracy of assembly, so as to speed up the tunneling speed and improve the safety of tunnel construction. Determining the position and pose of the target segment is the premise of realizing the automatic assembly of segments. At present, relevant researchers have proposed some shield segment positioning schemes based on vision or line laser sensors. However, these algorithms need to manually arrange the target on the shield segments or know the geometric size of the shield segments in engineering applications, which have high requirements for target layout and segment production. As a result, additional operation procedures and inevitable shield segment assembly errors are brought. Aiming at the problems of segment position and pose detection in the automatic shield segment assembly, a scheme of shield segment automatic assembly without marking segments in advance and suitable for unknown geometric parameters is proposed, and the segment position and pose detection algorithm is derived, which provides an effective technical solution for the segment position and pose detection in the automatic shield segment assembly engineering.

Methods The installation layout of five line laser sensors, which are utilized to measure the position and pose of shield segments and the segment assembly process are reasonably designed. Specifically, these sensors, installed on the grasping plate of the end effector of the assembly robot are used to illuminate the straight and arc edges of the segments respectively, and the 3D coordinates of the deflection points generated by the line laser irradiation on the segment edges are detected. The position and pose of the segments with respect to the base coordinate system of the assembly robot can be calculated gradually during the assembly process, and then the position and pose of the end effector of the robot with respect to the base coordinate system can be calculated when the robot assembles the segment to the desired pose, so as to realize the automatic assembly of the segments.

Results and Discussions In the laboratory environment, the line laser sensor (OPT-LPB400) and the 6-DOF series robot (INOVANCE-IRS311-7-70TS5) are used to build the experimental platform for automatic segment assembly. In order to verify the effectiveness of the segment position and pose detection algorithm, an automatic segment assembly experiment is carried out on the platform. The experimental results show that the proposed shield segment automatic assembly method shows a small error in segment positioning, and the maximum level difference of segment assembly is 1.50 mm, and the maximum gap is 0.96 mm, which meets the engineering requirements.

Conclusions An automatic shield segment assembly method based on five line laser sensors is proposed. This method does not need to mark the segments in advance, and the shield segments can be assembled automatically even when the geometric parameters of the shield segments are unknown, which is simple and reliable. The Savitzky-Golay (SG) filter algorithm is used to smooth the line laser profile data The automatic detection of segment edge point coordinates, based on curvature characteristics, is then realized, enabling the automatic detection of segment assembly pose. The segment assembly process is designed. The position and pose detection algorithm in the shield segment assembly process is derived according to the process. Only the three-dimensional coordinates of the deflection points illuminated by the line laser sensors on the edges of the shield segments were used to detect the position and pose of the shield segments, and then the desired assembly position and pose of the end-effector of the assembly robot were determined. The experiments of automatic segment assembly show that the proposed automatic shield segment assembly method realizes the automatic assembly of shield segments under laboratory conditions. The maximum level difference is 1.50 mm, and the maximum gap is 0.96 mm, which meets the engineering requirements of segment assembly. This method provides an effective technical solution for the automatic assembly of shield segments.