AngleFusion_dataset

This dataset provides synchronized measurements from resistive strain sensors and dual inertial measurement units (IMUs) mounted on the proximal and middle phalanges of the index finger for joint-angle estimation. Data were collected across five sessions (S11\u2013S15) under standardized flexion\u2013extension motion templates at a sampling rate of 100 Hz. Each CSV file contains aligned time-series of Euler angles (AngX, AngY, AngZ) from IMUs, strain-derived flexion angles, and ground-truth joint angles obtained from high-speed camera tracking. The dataset supports research on multimodal fusion, sensor calibration, drift\/hysteresis compensation, and real-time kinematic estimation for wearable hand-tracking systems. It was used to validate a causal temporal convolutional network (TCN) with modality-aware priors for stable, CPU-efficient inference. Researchers may use this dataset to benchmark sensor fusion, uncertainty modeling, or machine learning approaches for finger-joint angle regression and wearable rehabilitation applications.