error

Week Four Session Overview

  • Focus on 2D Kinematic Analysis

  • Recap of previous steps: data collection, recording, processing, and digitizing files

Understanding Noise in Data

  • Definition of Noise: Potential errors in data to either exclude or correct

  • Importance of data validity and reliability in biomechanics

Data Validity and Reliability

  • Validity: Reflects how accurately the data represents the variable being measured.

  • Reliability: Consistency of repeated measures.

Types of Errors in Data Collection

  • Systematic Errors

    • Consistent throughout the recording process.

    • Examples:

      • Inaccurate camera placement (not perpendicular).

      • Incorrect marker placement (not at joint centers).

      • Calibration errors (incorrect order or identification of points).

  • Random Errors

    • Affect specific frames or sections of data.

    • Examples:

      • Human error during digitizing (imprecise clicking).

      • Markers moving unexpectedly during the motion (e.g., using arms in a jump).

      • Low resolution or shutter speed affecting clarity in high-speed actions.

Effects of Errors on Data

  • Errors can lead to inaccurate tracking of joint centers.

  • Resulting data inaccuracies affect derived variables such as displacement and acceleration.

Importance of Accuracy and Precision

  • Accuracy: Being close to the true value.

  • Precision: Consistency of repeated measurements.

    • Possible scenarios:

      • High accuracy, high precision (ideal scenario)

      • Low accuracy, high precision (systematically wrong)

      • Low accuracy, low precision (inconsistent and wrong)

Measuring and Reporting Data Quality

  • Researchers often repeat measurements (5 times) to assess precision.

  • Importance of using proper equipment to ensure data integrity (camera type, recording frequency, resolution).

High Definition and Data Management

  • Utilizing HD resolution (e.g., 1920x1080) for clearer data collection.

  • Issues related to large file sizes and data storage requirements.

Systematic vs. Random Noise

  • Systematic Errors: Affect the entire recording, generally linked to accuracy.

  • Random Errors: Impact individual frames typically due to transient issues during recording/analysis, linked to precision.

Filters for Data Smoothing

  • General Concepts: Need to apply filters after collecting data to minimize random errors.

  • Examples of Filters:

    • Fourier Transformation: To identify and remove high-frequency noise.

    • Butterworth Filter: Effective for smoothing noisy data, used frequently.

    • Spline Filter: Often more accurate than traditional filters in preserving data integrity.

Considerations When Digitizing Data

/

  • Issues of filters introducing errors at the start/end of movements.

  • Importance of ignoring initial frames affected by filters in analysis.

  • Types of variables to export:

    • Linear variables: position, velocity, acceleration.

    • Angular variables: angle, angular velocity, acceleration.

Analyzing Kinematic Data

  • Kinematic data can be integrated with body mass for inverse dynamics calculations.

  • Impacts of improper digitizing include inaccuracies in force calculations.

  • Importance of minimizing analysis errors before presenting data.

Conclusion and Next Steps

  • Discussion on staying organized and proactive with group projects.

  • Homework task focusing on calculating the center of mass based on segment locations.