盾构姿态测量数据集

盾构机作为集机、电、液、信息、人工智能于一体的现代化隧道专用装备，可以同步完成开挖、推进、出渣和支护等工序，实现隧道一次成型。其具有掘进速度快、地质适应性强、安全性高、自动化程度高、劳动强度低、对施工场地影响小等优点，广泛用于水利水电隧洞、城市地铁、铁路公路隧道、矿藏巷道、城市管廊等洞室的开挖。国内外盾构隧道建设正呈现出向大深度、大断面、长距离方向发展的趋势。如何有效提升盾构施工智能化水平，从而提高施工工效和安全性已成为盾构施工领域亟待解决的关键问题。为解决上述问题，盾构推拼同步工法与自主掘进技术应运而生。在盾构自主掘进过程中，若盾构机与DTA产生偏差，纠偏轨迹规划是其完成掘进位姿偏差纠正的关键步骤。
惯性导航法，通过安装在盾构机上的惯性导航单元（IMU）测量角速率和加速度，积分得到盾构机的姿态、位置和速度。惯性导航单元具有可独立工作、安装空间小、测量速度快等优点，但其核心挑战在于惯性器件存在零偏及随机游走误差，导致姿态和位置信息随时间积累而发散，因此，在实际工程应用中，惯性导航系统必须引入激光靶等外部高精度测量方式对该误差进行实时估计与校正，形成组合导航系统。鉴于当前研究领域普遍缺乏具备完整性、精度可溯源性和高时空同步性的盾构位姿基准数据集，本数据集依托崇太长江隧道等重大工程的现场环境，结合实验室精密标定与第三方权威检测，构建了“盾构位姿高精度实时测量数据集”。本数据集验证本测量系统位姿测量的实时性、精确性和稳定性，为盾构的同步推拼提供精确、实时和可靠的盾构位姿。
本数据集涵盖系统运行的完整生命周期，其结构与重用价值体现在：它提供了高频现场实测数据，收集了惯性测量单元（IMU）与激光靶的原始观测数据（采样频率10Hz）及经融合后的位姿信息。其次，其引入的多维精度验证数据，如第三方高精度转台对比数据和系统测量周期报告，提供了可靠的“真值”参照和时延信息，用于精确评估融合算法的收敛性、稳健性以及系统在动态环境下的实时响应能力。此外，数据集包含的视觉参数标定数据，提供了大量的标定板图像和光斑扫描结果，相机内参模型和非线性畸变参数进行高精度优化，从而在底层感知层面矫正激光标靶的像素级测量误差，最终为位姿测量系统提供更精确、更低噪声的观测输入，以提升整体系统的误差估计与姿态矫正性能。最后，数据集提供的掘进轨迹闭环验证数据，整合了轴线规划数据与实测偏差数据，验证轨迹规划与轨迹纠偏控制策略的实际应用与验证。

As a modern dedicated tunnel equipment integrating machinery, electronics, hydraulics, information technology and artificial intelligence, the shield tunneling machine can synchronously perform excavation, propulsion, mucking and support processes to achieve one-pass tunnel forming. It has advantages such as high excavation speed, strong geological adaptability, high safety, high automation, low labor intensity and minimal impact on construction sites, and is widely used in the excavation of underground cavities including water conservancy and hydropower tunnels, urban subways, railway and highway tunnels, mineral roadways and urban utility tunnels. The construction of shield tunnels at home and abroad is showing a trend towards greater depth, larger cross-section and longer distance. How to effectively improve the intelligent level of shield tunneling construction to enhance construction efficiency and safety has become a key issue urgently to be solved in the shield tunneling field. To solve the above problems, the synchronous pushing and assembling method and autonomous tunneling technology for shields have emerged as the times require. During autonomous shield tunneling, if the shield machine deviates from the Design Tunnel Axis (DTA), deviation correction trajectory planning is a key step to correct the tunneling pose deviation.

The inertial navigation method measures angular rate and acceleration through an inertial measurement unit (IMU) installed on the shield machine, and integrates the measured data to obtain the pose, position and velocity of the shield machine. The IMU has advantages such as independent operation, small installation space and fast measurement speed, but its core challenge lies in the zero bias and random walk errors of inertial devices, which cause the pose and position information to diverge as time accumulates. Therefore, in practical engineering applications, the inertial navigation system must introduce external high-precision measurement methods such as laser targets to perform real-time estimation and correction of this error, forming an integrated navigation system.

In view of the general lack of shield pose benchmark datasets with completeness, traceable accuracy and high spatiotemporal synchronization in the current research field, this dataset is constructed based on the field environment of major projects such as the Chongtai Yangtze River Tunnel, combined with laboratory precision calibration and third-party authoritative testing, and is named "High-Precision Real-Time Measurement Dataset for Shield Pose". This dataset verifies the real-time performance, accuracy and stability of the pose measurement of the proposed measurement system, and provides accurate, real-time and reliable shield pose for the synchronous pushing and assembling of shields.

This dataset covers the entire life cycle of system operation, and its structure and reuse value are reflected in: First, it provides high-frequency field measured data, collecting original observation data from the inertial measurement unit (IMU) and laser targets (sampling frequency of 10 Hz) as well as fused pose information. Secondly, the introduced multi-dimensional accuracy verification data, such as third-party high-precision turntable comparison data and system measurement cycle reports, provide reliable "ground truth" references and time-delay information, which are used to accurately evaluate the convergence, robustness of the fusion algorithm and the real-time response capability of the system in dynamic environments. In addition, the visual parameter calibration data included in the dataset provides a large number of calibration board images and spot scanning results. The camera internal parameter model and nonlinear distortion parameters are optimized with high precision, thereby correcting the pixel-level measurement errors of the laser target at the underlying perception level, and ultimately providing more accurate and lower-noise observation inputs for the pose measurement system to improve the error estimation and attitude correction performance of the overall system. Finally, the tunneling trajectory closed-loop verification data provided by the dataset integrates axis planning data and measured deviation data, verifying the practical application and validation of trajectory planning and trajectory deviation correction control strategies.

搜集汇总

背景与挑战

背景概述

该数据集专注于盾构机姿态测量，旨在为智能盾构施工提供高精度、实时的位姿数据支持。它包含惯性测量单元（IMU）和激光靶的原始观测数据、融合后的位姿信息，以及多维精度验证数据，如第三方对比数据和视觉参数标定数据，用于评估算法性能和系统响应。数据集基于崇太长江隧道等工程现场构建，数据量1.22GB，覆盖完整系统生命周期，适用于盾构自主掘进、轨迹纠偏等关键技术研究。