Please enable JavaScript.

Coggle requires JavaScript to display documents.

Baca / Kornfeind, Own developments - Coggle Diagram

- - - - Kinematic parameters
        
        Instantaneous rate of change ...
        
        Angular acceleration: rate change angular velocity
        
        Linear displacement (r = m): change in position in space
        
        Angular velocity: rate change angular displacement
        
        Velocity (v = m/s): rate change displacement
        
        Angular displacement: change angular position
        
        Acceleration (a = m/s^2): rate change velocity
    - - Force = vector
        
        b = direction
        
        c = magnitude
        
        a = point of application
        
        weight, friction, air- & water resistance, muscle forces
      - change of motion / deformation
      - rotatory motions
        
        linear momentum = angular momentum
        
        L=J*w -> J smaller & w larger
        
        mass = moment of inertia
        
        J -> moment of torque = (M=J*a)
        
        force = moment of force
        
        muscle torque = (M=r*F)
    - - 2 = Law of momentum = (F=m*a) = acceleration direct proportional to the magnitude of force, inversely proportional to mass
      - 3 = Law of interaction = for every action on an object there ist an equal opposite reaction by the second object to the first
      - 1 = Law of inertia = object will stay in rest or linear motion unless external forces act
    - - point which moves as mass would be concentrated within it and the forces apply on it
      - object supported at CM = no net torque acting = static equilibrium
      - COM / CM
  - - - inertial sensors or IMU (inertial measurement units)
        
        combi of accelerometers, gyroscopes, magnetormeters
        
        rotations about 3 axes
        
        3D orientation
        
        Xsens, full-body suits/strap system (wireless & camera-less 3D)
        
        orientation of individual segments in 6 degrees of freedom (3 translational, 3 rotational)
      - orientation in space
      - opto-electronic motion analysis system
        
        real-time system
        
        video-based off-line system (Kinovea)
        
        contact-free & automatable
        
        active (LEDs) & passive markers (retroflective) -> protocols (e.g. Plug-in-Gait)
        
        calibration
        
        markerless systems
    - - force plates
      - pressure distribution
    - - geometric mass distribution
      - internal geometry of LS
      - linear measurements (length, circumference)
      - mechanical properties (density, etc.)
      - physical measurements of body
    - - temporal activation pattern of muscle
      - muscle activity
      - Contribution / non-contribution of a specific muscle to a motion
      - Timely coordination of muscles
      - Antagonistic / synergistic muscle activities
- - - - 3840x2160 (4K UHD)
      - 1920x1080 (1080p)
    - - px total = 1920 x 1080 = 2.073.600 ≈ 2 Megapixel = 1 frame
      - px total = 3840 x 2160 = 8.294.400 ≈ 8 Megapixel = 1 frame
    - - vertical field of view (FOV)
        
        depends also on qualitative vs quantitative video analyse
      - e.g. FOV = 180 cm, marker diameter = 2,5 cm
        
        720p enough (720x480px)
        
        FOV 1080p -> (1080px/5px)*2,5cm = 540cm
        
        (180cm/2,5cm)*5px = 360px
      - smallest object (marker) at least 5 pixels
      - simple lens 2D
      - simple lens 3D
      - image sensor (mm, inch) -> different size of image sensor & quality = increasing resolution = more memory needed
      - frame rate
        
        25/50/100 frames per seconds (fps) -> progressive vs interlaced
        
        1-3x object (marker) speed (cm/s) -> e.g. sprinter with 8-10 m/s >>> 1000 fps
      - video compression
        
        codecs: MPEG-2 , MPEG-4, H.264/H.265, WMV
        
        containers: avi, mp4, mov, mkv, fla, ...
        
        amount of video data
  - - - center of mass (COM)
      - overlay selected parameters
      - calculation of numeric values (distance, velocity, acceleration)
  - - - increase or decrease light falling on sensor
    - - how long light shines on sensor -> exposure time
      - 1/50s, 1/125s, 1/250s, 1/500s, 1/1000s -> available light
      - as low as possible (motion blur)
    - - manual focus -> avoids problems
      - autofocus -> needs time
    - - tripod
      - marker (joint center)
      - capture device
      - measuring tape (or calibration object)
      - -> Kinovea
  - - - graphics
      - time measurements
      - key frames (situations, phases)
      - text annotations
- - - - sensors / devices on subjects or sports equipment
      - data acquisition without influencing subjects
      - integrated application
      - local positioning measurement (LPM) system
    - - fitness studio
      - everyday's life
      - training -> adapt
      - motivation
      - care-taking
      - advice
    - - principle of operation
        
        electronic processing
        
        automatic detection of time intervals
        
        data transfer
        
        signal capturing (microphones)
        
        visualization of relevant information
  - - - monitoring activities
      - classification / quantification of activities
      - promoting PA
      - online-supervision
      - feedback
    - - GPS & ECG-signals usually valid
      - critical = PA from accelerometer data (influence of position of sensor; measures for better accuracy = lower acceptance)
      - wrist-worn devices -> good energy expenditure measurement
    - - ubiquitous computing / internet of things
      - wireless connection technologies -> miniaturization
      - collecting data
      - computer-based technologies on human body
      - intelligent wristbands, smartphones, smartwatches, smart textiles
    - - app-escape -> 80-90% deleted after 1 usage
      - overextension
      - lack of behavior change / gamification elements
  - - - statistical classification
        
        identification of the corresponding class of motion
        
        selection of appropriate feedback / adaptation
      - statistical decision
        
        decision engines
        
        bayes classifiers
        
        k-Nearest Neighbor (k-NN)
        
        binary classification trees
        
        rule based approaches
        
        linear discriminant classifiers
        
        support vector machines
      - unsupervised neural networks
      - specific models
        
        feedback, recommendations, ... based on prediction of parameter
        
        e.g. prognosis of HR during running -> recommendations for next