Skip to content

OUTLINE.rst

Latest commit

 

History

History
152 lines (108 loc) · 4.69 KB

OUTLINE.rst

File metadata and controls

152 lines (108 loc) · 4.69 KB

Journal

Jason would like to submit this to F1000Research as a data note.

http://f1000research.com/

See the Author Guidelines and check out the "Data Note" description:

http://f1000research.com/author-guidelines

Authors

This is the proposed author list, in order:

  • Jason K. Moore
  • Sandra K. Hnat
  • Antonie J. van den Bogert

Abstract

No longer than approximately 300 words; without references.

Introduction (optional)

Rationale for creating the dataset(s) and/or objectives for the experiment resulting in the dataset - why the data were gathered or produced.
  • System ID techniques require large amounts of rich data
  • Map human control into prosthetic control space
  • Identification of control mechanisms

Materials and methods

Detailed account of the protocol used to generate the dataset:

  • For standard protocols that have been published elsewhere, a brief description and reference is sufficient.
  • Details for the source of all samples, reagents, antibodies etc.
  • Details of how samples were selected; what exclusions were made, if any.
  • Details of what was being measured;

For processed data, this section should include details of any software used to process the data, including the version used, details of where the software can be accessed, and any parameters that could impact the outcome of the results.

Please see our editorial policies for patient data, or for research involving animals.

Equipment

  • Cortex 3 for motion capture (marker coordinate locations and GRF)
  • D.Flow 3.15-3.16 for VR program, treadmill manipulation, and data recording
  • Forcelink R-mill for the treadmill (+/- 0.05 meters mediolateral displacement and +/- 10 degrees sagittal pitch)
  • 10 Motion Analysis Osprey cameras

Protocol

  • describe the testing procedure for each subject and how that was handled with event timing in D-Flow
  • describe briefly in words the D-Flow program and what it's doing
  • 47 marker set, maybe include diagram in supplementary material
  • harness on subject, etc.

Perturbation Design

  • Simulink diagrams and MATLAB code
    • twice integrated white noise into high pass Butterworth and saturated
  • signals generated around ~8-10% standard dev. of the average speed
    • done ad hoc through coding, also done "experimentally"

"Raw" Data

  • The raw data will be defined as the data which comes out D-Flow along with the input belt signal generated by the Simulink model.
  • Data:
    • Marker locations in 3D space
    • Belt speed: actual and desired
    • Ground reaction loads at each foot
    • Events
    • Anatomical marker locations
    • Subject data: height, weight, age, ankle/knee widths
  • Detailed description of the meta data file.
  • File descriptions: mocap file, record module file, etc.

"Processed Data"

  • We will present some details of the processed data, but point the readers to the GATK for processing the raw data.
  • The processed data will be the outputs of the WalkingData class in the gait analysis toolkit.
    • Joint angles and angular rates
    • Joint torques
    • Heelstrikes and toeoffs
  • Description of GaitAnalysis Toolkit
  • All signals filtered at 6 Hz low-pass 2nd Butterworth filter (forward-backward)
  • Joint angles and moments calculated from 2D inverse dynamics program
  • Compensation Techniques (maybe)
    • inertial artifacts due to platform movement
    • did not use platform movement, but implemented in code because we intended to move the platform
    • belt acceleration compensation. This may be necessary.
  • Show typical plots of data.
    • one subject, medium walking speed, sagittal plane joint moments and angles in hip, knee, ankle (a 3x2 matrix of graphs). From the unperturbed walking a mean and SD, from the perturbed walking, maybe a 10-second section, or multiple gait cycles superimposed on a plot with 0-100% on the horizontal axis.
  • Pin to a particular version of GaitTK and DynamicistTK.

Dataset validation (optional)

Information about any validation carried out and/or any limitations of the datasets, including any allowances made for controlling bias or unwanted sources of variability.
  • Walker ?#9? has odd ankle joint torques (trials 25, 26, 27)

Data availability

A machine-readable section which will make it possible for the citation and provenance of the dataset(s) to be tracked; this section will be generated by the editorial office.

I'd like to share the data on Figshare. I'm not sure if they support a directory hierarchy and they don't support uploading giant compressed files (unless maybe if you pay).