IMEB-Exo Synthetic Dataset: Time-Series Biomechanical Data and Multicriteria Indicators for Passive Upper-Limb Exoskeleton Assessment

The IMEB-Exo Synthetic Benchmark Dataset is an open and reproducible dataset developed for the analysis of passive upper-limb exoskeletons, occupational biomechanics, human–exoskeleton interaction, and multicriteria biomechanical decision support. It provides deterministic synthetic time-series data and derived biomechanical indicators designed to support methodological validation, sensitivity analysis, benchmarking, and reproducibility of the IMEB-Exo framework (Multicriteria Biomechanical Exposure Index for Exoskeletons).

The dataset represents 24 virtual anthropometric models evaluated under three simulated occupational task scenarios and two experimental conditions: without exoskeleton assistance and with a passive upper-limb exoskeleton. The data structure enables comparisons between assisted and non-assisted conditions across multiple anthropometric profiles, task scenarios, and human–exoskeleton interface regions.

The time-series dataset includes normal and tangential human–exoskeleton interface forces, estimated interface pressure when a technically defined contact area is available, shoulder and trunk kinematics, exoskeleton assistance torque, joint reaction variables, muscle activation surrogates, and cumulative biomechanical exposure indicators.

Derived variables include peak and mean force, 95th-percentile force, root mean square (RMS) force, force–time integral, loading rate, time above an operational load threshold, peak counts, projected occupational exposure, CRITIC objective weights, TOPSIS results, IMEB-Exo scores, biomechanical criticality classes (CCB-Exo), and rankings of human–exoskeleton interface regions.

The repository contains:

• deterministic synthetic biomechanical time-series data;
• interface-level and scenario-level metrics;
• IMEB-Exo scores and regional rankings;
• CRITIC objective weights and TOPSIS outputs;
• metadata for 24 virtual anthropometric models;
• task and scenario definitions;
• human–exoskeleton interface definitions and contact-area specifications;
• generation parameters;
• a structured SQLite database;
• CSV and compressed CSV datasets;
• Python scripts for dataset generation and IMEB-Exo calculation;
• data dictionary and schema documentation;
• SHA-256 checksums and file inventory;
• representative figures and data excerpts;
• reproducibility documentation in English and Portuguese.

The dataset is intended for researchers working with occupational biomechanics, passive occupational exoskeletons, wearable robotics, ergonomic risk assessment, musculoskeletal modelling, multicriteria decision analysis (MCDA/MCDM), decision-support systems, synthetic biomechanical datasets, and reproducible computational research.

Methodological notice: the dataset is synthetic and deterministic. The generated variables reproduce the expected structure and behaviour of an OpenSim-oriented musculoskeletal simulation workflow but do not constitute experimentally measured human data and must not be interpreted as direct outputs from an executed OpenSim simulation. The dataset should therefore be used as a methodological benchmark for algorithm development, reproducibility testing, sensitivity analysis, multicriteria modelling, and computational workflow validation rather than as experimental evidence of human biomechanical responses.

The dataset was developed within the project “Multicriteria Methods for the Analysis of Monitored Data in the Use of Passive Upper-Limb Exoskeletons”, supported by FAEPEX / University of Campinas (UNICAMP), Grant No. 70262-26.