Wearable Sensors in Exercise and Sport Science

Learning Objectives for Wearable Sensors

• LO1: Understand principles of data collection for biomechanics, especially with regard to the principles of filming movements for quantitative analysis.
• LO2: Understand how the mechanical properties of biological tissues influence the response of the body to these loads, potentially causing acute and chronic injuries.
• LO4: Be able to assess the demands placed on the body by exercise loads and use this to recommend changes to enhance performance and reduce injury risk.
• LO6: Be able to conduct a biomechanical assessment of movement technique and communicate the findings to a lay audience.

Inertial Measurement Units (IMU) Fundamentals

• Definition of an IMU: An Inertial Measurement Unit is a device that often houses multiple sensors to track movement. These typically include:

  • Accelerometers.
  • Gyroscopes.
  • Magnetometers.
  • Global or local positioning systems (GPS/LPS).

• Mechanism of Accelerometers:

  • Operation is based on Newton’s Law of Acceleration: $F = m \times a$.
  • The signal voltage produced is proportional to the acceleration ($a$).
  • The internal mechanism involves a mass accelerating against a force transducer within a housing.
  • Force Transducers: Common types include strain gauges or piezoelectric sensors.
  • Triaxial Accelerometers: These consist of three individual accelerometers oriented at right angles to one another ($x$, $y$, and $z$ axes).

• Advantages of Accelerometers:

  • Signal output is available immediately for real-time analysis.

• Disadvantages of Accelerometers:

  • Relative Acceleration: Measures acceleration relative to its specific position on a body segment, not necessarily the center of mass.
  • Shock Sensitivity: Highly sensitive to mechanical shocks.
  • Movement Artefact: Susceptible to noise created by the sensor moving relative to the skin or clothing.

Market Comparison of Wearable Sensors

• Blue Trident (Vicon):

  • Features: Tri-axial accelerometer (Low: $\pm 16g$ at $1125\,Hz$; High: $\pm 200g$ at $1600\,Hz$), Gyroscope ($1125\,Hz$, $\pm 2000^{\circ}/s$), Magnetometer ($\pm 4900\mu T$ at $112\,Hz$ or $70\,Hz$ for global angles).
  • Primary Uses: Sport and research.

• GENEActiv:

  • Features: Tri-axial accelerometer ($\pm 8g$ at $10-100\,Hz$).
  • Primary Uses: Physical activity intensity, posture changes, and sleep tracking.

• ActiGraph (Ametris):

  • Features: Tri-axial accelerometer ($\pm 8g$ at $32-256\,Hz$), Gyroscope ($\pm 2000^{\circ}/s$), Magnetometer.
  • Primary Uses: Physical activity intensity, clinical studies, and sleep.

• Consumer Smartwatches (Apple, Fitbit, Garmin):

  • Apple Watch: Accelerometer ($\pm 8g-16g$ at $50-100\,Hz$), Magnetometer. Used for fall detection and activity.
  • Fitbit Watch: Accelerometer ($\pm 2g-8g$ at $25-100\,Hz$). Used for activity and sleep.
  • Garmin Watch: Accelerometer ($\pm 8g-16g$ at $25\,Hz$), Magnetometer. Used for activity and sleep.

Specific Applications: IMeasureU and IMU Step

• IMU Step (Blue Trident): Utilizes shank-mounted accelerometers to measure step impacts.
• Step Intensity Categorization:
  • Low Intensity: $2-4G$.
  • Moderate Intensity: $4-10G$.
  • High Intensity: $10-14+G$.
• Sport Applications: Running, basketball, kicking, tennis, cricket.
• Value Proposition: Monitoring lower limb loads for injury rehabilitation by merging intuition with quantitative data.

Critiquing Metric Definitions: PlayerLoad

• Terminology Errors: The term "Load" implies a kinetic variable (involving force). However, acceleration is a kinematic variable (describing motion without respect to force), making the term technically incorrect.
• Definition Issues:
  • Jerk: This is defined as the rate of change of acceleration.
  • It is the derivative of acceleration with respect to time: $\frac{da}{dt}$.
  • It is the third derivative of position with respect to time: $\frac{d^3x}{dt^3}$.
  • Mathematical Ambiguity: The standard "division of the sum of 100" used in commercial PlayerLoad formulas has not been clearly scientifically defined.

Cautions in Using Triaxial Accelerometers (Edwards et al., 2018)

• Location of Measurement: Sensor location often has poor reliability and may not provide a valid representation of thoracic segment vertical acceleration.
• Source of Measurement Error: Elasticised harnesses in player tracking units are a substantial factor in overestimating peak vertical acceleration during running.
• Variable Accuracy: Trunk-mounted units (such as GPSports) are often unable to accurately estimate peak vertical Ground Reaction Force (vGRF) or 3D center of gravity acceleration.

Specialized IMU Components: Magnetometers and Gyroscopes

• Magnetometer:

  • Function: Measures the direction and magnitude of external magnetic fields.
  • Utility: Provides an external reference to the horizontal plane and can determine trunk inclination in static positions.
  • Limitations: Highly affected by metal (e.g., reinforced concrete). Accuracy varies by geographic location; it is best at the equator and poor near the poles.

• Gyroscope:

  • Function: Measures angular velocity by sensing Coriolis acceleration.
  • Governing Formula: $f_c = 2m \times v \times \omega$.
    • $f_c$ = Coriolis force.
    • $m$ = vibrating mass.
    • $v$ = instantaneous linear velocity of the mass (caused by the actuator).
    • $\omega$ = angular velocity of the sensor.
  • Measurement: Units measure roll, pitch/yaw, and turn rate. The measured force is directly proportional to the angular velocity.

Positioning Systems: Global (GPS) vs. Local (LPS)

• Global Positioning System (GPS): Uses a minimum of 3 satellites to determine the 3D position of an object globally.

• Local Positioning System (LPS): Uses a set of signaling beacons to calculate 3D position within a specific local field, typically used indoors where GPS signals are unavailable.

• Device Comparison (LPS/GPS):

  • KINEXON (Local): $20\,Hz$ LPS, $1000\,Hz$ Accelerometer ($\pm 16G$), $100\,Hz$ Magnetometer, $200\,Hz$ Gyroscope ($4000^{\circ}/s$).
  • Catapult Vector Pro T7 (Local): Up to $100\,Hz$ LPS.
  • Catapult Vector Pro S7 (Local + Global): $10\,Hz$ GPS and up to $100\,Hz$ LPS.
  • STATSports Apex Pro (Global): $10$ or $18\,Hz$ GPS.

Accuracy and Reliability of Positioning Systems

• ClearSky Catapult (Indoor LPS):

  • Mean difference of $0.21 \pm 0.13\,m$ compared to Qualisys motion capture (gold standard).
  • Low errors observed for position, distance, and average speed.
  • Large errors observed for instantaneous speed (the largest difference between LPS and 3D motion capture).
  • Reliability is highly dependent on the placement of nodes.

• Factors Affecting Signal Quality:

  • Environmental obstructions and geographic location.
  • Number of satellites: Minimum of 3 required, but ideally 6 or more.
  • Horizontal Dilution of Precision (HDOP): Represents the accuracy of the horizontal positional signal based on the geometrical organization of the satellites.

Calculating Velocity and Acceleration in GPS

• Positional Differentiation: Calculating distance based on changes in location between signals. This tracks distance and derived velocity.
• Doppler Shift: Measuring the periodic signal emitted by satellites to obtain an almost instantaneous measure of velocity.
• Calculations:
  • $Distance = Velocity \times Time$.
  • Acceleration in $10\,Hz$ GPS is typically calculated over a window of $0.2$ or $0.3\,s$.

The "Black Box" of Data Processing

• Manufacturer Algorithms: Software often automatically interpolates, smooths, or extracts data using hidden (black box) algorithms.
• Minimum Effort Duration: There is no consensus on the duration required to define an effort. For example, a sprint might be defined as a speed $\ge 7\,m/s$ maintained for a minimum of $0.4\,s$ (4 consecutive samples at $10\,Hz$).

Speed Zones and Sport Standardization

• Lack of Standardization: Speed zones vary significantly between sports (Rugby, Soccer, AFL) and within the same sport based on sex or elite status.
• Arbitrary Zone Examples:
  • Rugby Union (Elite Females): Zone 1 Walking ($<2\,m/s$), Zone 5 High-intensity ($>5\,m/s$).
  • Soccer (Elite Males): Sprinting often defined as $>5.31\,m/s$.
  • Australian Rules Football: High-speed running often defined as $4.17-10\,m/s$.
  • Hockey: High-speed defined as $>4.17\,m/s$.

Practical Application for Practitioners

• Validity Benchmarks: Practitioners should compare wearable data against gold standards: 3D Optoelectronic motion capture, timing gates, or radar/laser guns.
• Decision Making: Accuracy and precision must be known when:
  1. Determining which specific metrics to track.
  2. Progressing or regressing an individual’s training load.
  3. Providing "top up" sessions by comparing actual loads to planned loads.