I

week 1

Schedule overview

  • Week 1 (commencing 24/7/23): Motion Analysis (1); Lab: 2D Motion Analysis - High-speed video (I-Speed / iPad) and Radar Gun
  • Week 2 (31/7/23): Measurements of forces and movement (2); Lab: Force
  • Week 3 (7/8/23): Electromyography (3); Lab: EMG
  • Week 4 (14/8/23): Data filtering (4); Lab: 3D Motion Analysis - Vicon motion capture
  • Week 5 (21/8/23): Free Body Diagrams (5); Lab: Joint torque & free body diagrams
  • Week 6 (28/8/23): Quantitative biomechanics applications (6); Lab: Writing skills and quantitative analysis proposal
  • Week 7 (4/9/23): A2: Mid-semester test; Assessment 2: Mid-Semester Test in seminar time; Assignment work
  • Week 8 (11/9/23): How do we move? (Use of muscles, tendons and nervous system in movement) (8); Assessment of SSC
  • Week 9 (18/9/23): What robot do we need? (9); Lab: Isokinetic dynamometry
  • 25/9/23: Mid-semester Break
  • Week 10 (2/10/23): How do we build the robot? (Effect of training on muscle, tendon and nervous system) (10); Applications in strength and conditioning - Assessing Power with Force Platforms
  • Week 11 (9/10/23): Footwear biomechanics (11); Technology applications 1
  • Week 12 (16/10/23): Developmental biomechanics and Occupational biomechanics (12); Technology applications 2
  • Week 13 (23/10/23): Revision; Technology applications 3; A1: Assignment due; Assessment 1: Assignment due 30/10/23; Study Week
  • Study Week (30/10/23); Exams (6/11/23) with Assessment 3: Exam (13/11/23); Exams; Deferred Exams (4/12/23)

Module 1 - Introduction to Quantitative Analysis

  • Core topics: Motion Analysis, Measurements of Forces and Movement, Electromyography (EMG), Data Filtering in Biomechanics, Free Body Diagrams, Quantitative Biomechanics Applications
  • Week 1 Agenda: Module 1 – Focus on ‘quantitative’ analysis; Example shown: Raw EMG Data trace (e.g., -2000, -1500, -1000, -500, 0, 500, 1000, 1500, 2000 mV) from Clark & Weyand (2014)

Concept: Objective of biomechanical analysis (example scenario)

  • Scenario: Long jumper analysed by biomechanists (Coach S&C Biomech Physiologist Physio Doctor)
  • Objective: Analyse technique during a jumping-type movement pattern
  • Focus measures: Knee joint moments/torques
  • Compare conditions: Injured vs Uninjured
  • Key relation: Moment = Force × Moment Arm
  • Moment is dependent on:
    • Magnitude of force
    • Angle of application
    • Length of moment arm

Data acquisition and workflow

  • Data collection/reporting pipeline:
    • Report
    • Smooth-calculate
    • Film subject
    • Digitise
  • Cameras and data types:
    • Types of Cameras
    • 2-D or 3-D?
    • Calibration
    • Analysing the Movement
    • Data Smoothing
  • Reminder: If displacement or distance can be quantified and the time period between pictures is known, then we can calculate the derivatives (velocity and acceleration):
    • Velocity: v = rac{\Delta s}{\Delta t}
    • Acceleration: a = \frac{\Delta v}{\Delta t}

Types of cameras and sampling considerations

  • Types of cameras:
    • High Speed Video: Olympus I-Speed
    • Opto-Electronic: Vicon, Motion Analysis
    • 2D Video Cameras: standard 2-D video setups
  • 2-D vs 3-D considerations:
    • Lines of resolution (e.g., 625 lines for PAL/NTSC)
    • Standard frame rate: 25/30 frames per second (PAL/NTSC format)
    • Image size, zoom, still imaging, panning
    • Shutter speed: faster shutter speed = less light but reduced blur
    • High shutter speed example: 1/500 s
  • High Speed Video features:
    • Very high frame rates (e.g., >30,000 frames per second)
    • Critical for identifying precise impact events
    • Expensive

Video frames, fields and sampling considerations

  • PAL/NTSC frame vs field structure:
    • Video frames are made up of 2 interlaced fields (Field A and Field B)
    • For PAL: there is 1/25 s between frames and 1/50 s between fields
  • Implication:
    • High-speed capture needed for very fast events (e.g., impact) due to temporal resolution requirements

Nyquist–Shannon sampling theorem and camera choice

  • Core idea: Must sample at least twice the highest frequency in the signal to avoid aliasing
    • If sprint cycling ≈ 160 rpm = 2.67 Hz, then required sampling rate f_s ≥ 2 × 2.67 ≈ 5.33 Hz
    • Normal system of 25 Hz easily satisfies this for slower movements; high-speed cameras may be needed for very rapid events (impact, fast throws, etc.)
  • Practical takeaway:
    • For many lab motions, standard 25 Hz suffices, but select higher sampling rates for fast events

2-D analysis: advantages and disadvantages

  • Positive aspects:
    • Cheaper and easier to use; fewer cameras required
    • Conceptually simpler; preselected movement plane
  • Limitations:
    • Rotational movements or movements out of plane introduce error
    • Perspective error: apparent size changes as object moves closer/farther from camera
    • Parallax error: apparent size changes as object moves across the camera view
  • 2-D analysis specifics:
    • Example resolution: 768 horizontal pixels (PAL) × 576 vertical pixels (PAL)
    • 2-D analysis often used for ball release or other planar motions

Parallax error in 2-D analysis

  • Parallax error example: Same athlete, same frame, different camera angle yields different results
  • Devices like iPads/phones offer convenient video but come with exposure to perspective limitations

Opto-electronic 3-D motion analysis

  • 3-D motion capture workflow:
    • Place retroreflective markers on the subject
    • Typical systems record up to ~500 frames per second
    • Best accuracy but requires expensive equipment
  • Decision: 2-D vs 3-D analysis depends on the movement and required detail
  • Core idea: 3-D analysis provides a more accurate representation of multi-planar movements

3-D motion analysis: calibration and methods

  • Calibration overview:
    • Wand calibration (Bartlett 1997): 3-D calibration uses a T-shaped object with known points moved in front of all cameras
    • In-built algorithm performs calibration
  • Calibration goals:
    • Map camera images to actual 3-D space
    • Establish coordinate correspondences across cameras
  • Direct Linear Transformation (DLT):
    • A method used to compute 3-D coordinates from multi-camera images
    • Requires at least two cameras, preferably ~45° apart; ECU system uses 10+ cameras

Data smoothing and derivatives in biomechanics

  • Before calculating velocity and acceleration, data must be smoothed to reduce error
  • Smoothing methods include:
    • Digital filters
    • Various spline techniques
  • Kinematic equations (from the session):
    • Velocity: v = \frac{\Delta s}{\Delta t}
    • Acceleration: a = \frac{\Delta v}{\Delta t}

2-D analysis – parameter identification and measurement (example parameters)

  • 2-D analysis considerations (ball-related measurements):
    • Height at set position
    • Height of the ball just prior to knee bend (max ball height)
    • Height at ball release
    • Height at max ball height (during flight)
    • Angle of arm at ball release
    • Angle of upper arm relative to forearm at ball release (arm straight = 180°, flexed = 90°)
    • Resultant velocity of ball at release (derived from components)
    • Release angle of ball (θ) = atan(Vv / Vh)
  • Key note: 2-D analysis identifies parameters to measure; 500 pixels ≈ 1 meter (calibration example from the slide)

2-D calibration example (netball analysis and an applied calculation)

  • Calibration idea: If 500 pixels correspond to 1 meter, and a runner travels 200 pixels between successive frames, real distance = (200 / 500) × 1 m = 0.4 m
  • Calculation steps shown in the slide:
    • Distance = (Pixels / Calibration Pixels) × Calibration Distance
    • With 200 pixels and 500 calibration pixels: Distance = (200/500) × 1 m = 0.4 m
  • Velocity example using frame rate:
    • Frame rate = 25 fps ⇒ 1 frame time = 1/25 s = 0.04 s
    • If travelled = 0.4 m in 1 frame, then velocity = 0.4 / 0.04 = 10 m/s

3-D motion analysis: calibration procedure details

  • 3-D calibration (Bartlett 1997, p. 185): wand calibration
  • Wand calibration steps:
    • Use a T-shaped object with known points moved in front of all cameras
    • System computes calibration automatically via a built-in algorithm
  • Result: 3-D reconstruction with multi-camera data

Break Motion Analysis in Biomechanics – key topics recap

  • Core components:
    • Types of Cameras (2-D vs 3-D)
    • Calibration procedures
    • Analyzing the Movement
    • Data Smoothing
  • Digitising for quantitative kinematic analysis:
    • Identification of landmark positions (usually joints)
    • Each landmark is digitised to give a coordinate in a digital space
    • Applicable to training and competition analysis

Analysing the Movement – 2-D specifics

  • 2-D analysis setup (example):
    • Resolution: 768 horizontal pixels (PAL) × 576 vertical pixels (PAL)
  • 2-D analysis pipeline: digitising selected landmarks into a biomechanical software
  • Examples of measured parameter sets include ball release-related metrics

Analysing the Movement – 3-D specifics

  • 3-D data creation relies on multi-camera capture and 3-D reconstruction techniques like DLT
  • Key advantage: closer to real, multi-planar movement representation
  • Additional considerations: calibration accuracy, time synchronization across cameras, and computational load

Practical considerations and implications

  • Pros and cons of camera choices:
    • 2-D cameras are cheap and easy but may introduce perspective/parallax errors for out-of-plane motions
    • 3-D systems offer higher accuracy for complex motions but are expensive and complex
  • Data management:
    • High frame rates yield large data volumes; storage and processing time are important considerations
  • Light requirements:
    • Higher shutter speeds require more light
  • Real-world relevance:
    • Applications include sports performance optimization, rehabilitation, footwear biomechanics, and occupational biomechanics
  • Ethical/practical implications:
    • Privacy and consent for video capture in training/competition settings
    • Handling and interpretation of data to inform training decisions responsibly

Key formulas and notation (summary)

  • Moment: M = F \times r where F is force and r is the moment arm length
  • Kinematic derivatives: velocity and acceleration
    • v = \frac{\Delta s}{\Delta t}
    • a = \frac{\Delta v}{\Delta t}
  • 2-D ball release parameters:
    • Release angle: \theta = \arctan\left(\frac{Vv}{Vh}\right) where Vv and Vh are vertical and horizontal velocity components
  • Calibration and scaling (2-D):
    • Scale Factor: \text{Scale Factor} = \frac{\text{Real-world length}}{\text{Pixel length}}
    • Actual Length: \text{Actual Length} = \text{Pixel length} \times \text{Scale Factor}
    • Distance: D = \left( \frac{\text{Pixels}}{\text{Calibration Pixels}} \right) \times \text{Calibration Distance}
  • Frame timing: \Delta t = \frac{1}{f_s} (e.g., for 25 Hz, \Delta t = \frac{1}{25} = 0.04\,\text{s})
  • Interlaced frame/field timing (PAL): between frames \approx 1/25\,\text{s} and between fields \approx 1/50\,\text{s}

Notable equipment mentioned

  • Cameras: Olympus I-Speed (High-Speed Video); Vicon; Motion Analysis (Opto-Electronic) systems
  • Frame rates and formats: PAL/NTSC; standard video: 25/30 fps; high-speed: up to thousands of fps

Assessment and study cues

  • A1: Assignment due 30/10/23
  • A2: Mid-semester test in seminar time
  • Exam schedule: 6/11/23 and 13/11/23; Deferred Exams 4/12/23
  • Emphasis on understanding the data pipeline from data capture to analysis (digitising, calibration, smoothing, derivative calculation) and on selecting appropriate 2-D vs 3-D approaches depending on movement complexity and accuracy needs