Knee-Exoskeleton-Data：Dataset for biological joint moment estimation of knee exoskeleton control
Description
This dataset is designed specifically for developing task independent knee exoskeleton control algorithms, which estimate biological joint torques from a joint perspective through deep learning. The dataset consists of two parts: the training set and the validation set, covering 11 motion modes closely related to the knee joint (flat walking, slope walking, obstacle crossing walking, random walking, weight-bearing walking, circuitous walking, reverse walking, going up and down stairs, squatting, sitting standing transition, and start stop), aiming to support the development and evaluation of a universal exoskeleton control model that does not require task recognition. The data file types for each motion mode are CSV format that can be directly used for model training. There are a total of 7 columns in the CSV file, including timestamps (seconds), left and right hip joint angles (degrees), left and right knee joint angles (degrees), and left and right knee joint bio torques (Nm/kg).
Files
Institutions
- Zhejiang Sci-Tech UniversityZhejiang, Hangzhou