Human MoCap data

The operator open access data provided in this study includes, first and foremost, a collection of isolated recordings of basic actions performed by the operator. These actions were captured using motion capture technology in a controlled and safe laboratory environment. The action segments represent elementary movements or routine tasks typically carried out during plant operations, for example, taking a step forward, kneeling, executing a shutdown from the control room, or inspecting a sensor. By concatenating these elementary recordings, it is possible to reconstruct complete operative scenarios. This modular approach allows for flexible simulation of different workflows by assembling sequences of basic actions, provided the layout of the plant and the relevant points of interest are known. Such flexibility supports the creation of fully digital yet realistic operator routines for further analysis or testing. The data are structured in tabular form: in the sheet named as the file, each row corresponds to a frame and includes, in order, the frame number, timestamp, and the (x, y, z) coordinate triplets for each sensor. Sensor names are listed in the header for ease of identification. This structure enables continuous spatial tracking of the operator throughout the duration of the recordings. As for “Sheet1” in each file, here the data represent the “deltas”, namely the variation in terms of position between two consecutive frames for each coordinate of each sensor. In this way, it is possible to define how each sensor moves over time, drawing a “trajectory” that identifies each basic action’s movement and enabling the concatenation between actions. In addition, the provided data also includes full recordings of real operational scenarios captured directly on field using the same motion capture technology, including recordings in which the plant shutdown was carried out from the control panel rather than the control room, as is typically the case.本研究提供的开源操作员数据，首要包含操作员执行基础动作的独立录制片段集。上述动作于受控安全的实验室环境中，通过动作捕捉技术（motion capture technology）采集。这些动作片段涵盖工厂运维作业中典型的基础动作或日常任务，例如向前迈步、下跪、从控制室执行停机操作，或是对传感器进行巡检。通过拼接这些基础录制片段，可重构完整的运维场景。该模块化方案支持通过组装基础动作序列，灵活模拟各类工作流程——前提是已知工厂布局与相关关注点位。此灵活性可助力构建兼具数字化与真实感的操作员操作流程，用于后续分析或测试。 数据以表格形式组织：在以文件名命名的工作表中，每一行对应一帧数据，依次包含帧编号、时间戳，以及各传感器的(x, y, z)坐标三元组。传感器名称列于表头，便于识别。此结构可实现录制全程中操作员的连续空间追踪。至于各文件中的“Sheet1”工作表，其数据代表“增量值”，即各传感器每个坐标在相邻两帧间的位置变化量。借此可明确各传感器随时间的运动情况，绘制出标识各基础动作运动轨迹的曲线，同时实现动作间的拼接。 此外，本次提供的数据还包含采用同款动作捕捉技术，直接在实地采集的真实运维场景完整录制片段，其中涵盖从控制面板而非常规控制室执行工厂停机操作的录制内容。