CLAMP: CLassified and Annotated Multimodal Postures.

CLAMP (CLassified and Annotated Multimodal Postures) is a multimodal dataset for hand posture recognition and grasp-phase analysis. It contains synchronized surface electromyography, accelerometer, and gyroscope recordings acquired with a Myo armband during predefined hand postures and grasp-related tasks. The dataset includes both raw MATLAB recordings and processed data organized by modality and grouping strategy. Processed files include signal matrices, MATLAB-compatible bundles, trial-level descriptors, sample-level annotations, and metadata. These annotations enable posture classification, multimodal sensor fusion, and temporal analysis of grasp onset, grasp completion, and grasp-related phases. The dataset is intended for research in wearable sensing, sEMG-based classification, human–machine interaction, low-cost hand control, and assistive robotic systems. Subject identifiers are anonymized, and no directly identifying personal information is included. If you use this dataset, please cite the associated Data in Brief article and the related Sensors 2024 publication.