The doi above is for this specific version of this dataset, which is currently the latest. Newer versions may be published in the future. For a link that will always point to the latest version, please use
doi: 10.4121/20418720

Schulte, Robert (2023): Roessingh Research & Development-MyLeg database for activity prediction (MyPredict). Version 1. 4TU.ResearchData. dataset. https://doi.org/10.4121/20418720.v1

Dataset

• Biomedical Engineering
• Human Movement and Sports Science
• Engineering
• Medical and Health Sciences

activity recognition, EMG, kinematics, lower limb, machine learning

CC BY 4.0

RefWorks, BibTeX, Reference Manager, Endnote, DataCite, NLM, DC, CFF

by Robert Schulte

Roessingh Research & Development-MyLeg database for activity prediction (MyPredict). The general aim of this database is to promote research in data-driven intent recognition strategies and activity prediction strategies in the lower-limb using electromyography and to promote research and development in the area of multi-array sEMG in the lower limb. The database contains three data sets, each containing kinematics and sEMG from able-bodied subjects. In total 55 subjects participated over 85 measurement sessions. Each data set contained a different sEMG measuring protocol containing either traditional bipolar sEMG or multi-array sEMG or a combination of both. In these data sets the subjects transitioned freely from one activity to the next, providing challenging data sets for activity recognition and providing the possibility to investigate human kinematics and sEMG during gait-related activities. This dataset is described in detail in Database of lower limb kinematics and electromyography during gait-related activities in able-bodied subjects (Schulte et al.)

MyPredict consists of three datasets, denoted by MP1XX, MP2XX and MP3XX.

• MyPredict 1: MP101-MP110, 10 able-bodied subjects (sex: 7m, 3f; age: 24±2 years; weight: 77±10 kg; height: 183±9cm), measured once
• MyPredict 2: MP201-MP235, 35 able-bodied subjects (sex: 14m, 21f; age: 23±2 years; weight: 73±11 kg; height: 179±9 cm), measured once
• MyPredict 3: MP301-MP310, 10 able-bodied subjects (sex: 4m, 6f; age: 24±2 years; weight: 71±9 kg; height: 174±6 cm), measured 4 times

These files contain the measurement moment named `Day_X' with X the number of the measurement moment. Inside these measurement moments there are files called `Trial_YY', with YY the trial number, containing the different data types and `MVC' containing the EMG maximum voluntary contractions of each measurement moment. Note that only MyPredict 3 contains multiple measurement moments per subject.

The different data types are acceleration (Acc), angular velocity (Gyr), joint angles (Ang), Orientation (Ori) and electromyography (EMG). Inside each file there are trials containing data arrays with the corresponding data. Data arrays are named as follows: Type_Side_Loc. Type is one of the six data types, Loc is the location of the sensor and Side is the side of the location, either Left, Right or empty. For example Ang_Right_Knee contains the 3D joint angles of the knee, Gyr_Pelvis contains the 3D angular velocity of the pelvis IMU and EMG_Left_VL contains the EMG data of the left vastus lateralis. Orientation is the orientation of the pelvis in space, expressed in Euler angles. Separate data types are 'Labels', which contains manual placed activity labels for each timestamp and 'Time' which indicates the timestamps per file. Marker data (Mrk) are stored in a separate group, `Markers' with their own `Time' array, as they have a different sample frequency (100Hz) compared to the other data types (1000Hz).

Code supporting this dataset can be found in the github repository: github.com/Rvs94/MyPredict

• 2023-05-26 first online, published, posted

4TU.ResearchData

*.HDF5

• https://doi.org/10.3389/frobt.2021.710806
• https://doi.org/10.3389/frobt.2022.869476
• https://doi.org/10.3390/s22176351
• https://doi.org/10.3990/1.9789036554862
• https://github.com/Rvs94/MyPredict

Roessingh Research & Development, Enschede, The Netherlands
University of Twente, Faculty of Electrical Engineering, Mathematics and Computer Science, Biomedical Signals and Systems