Thorsten Dahmen, Stephan Mantler · 2016. 3. 30. · Thorsten Dahmen, Stephan Mantler 1 Project...

1
Explorative Analysis and Visualization of Large Information Spaces Realistic Simulation and Optimization of Race Bike Training Thorsten Dahmen, Stephan Mantler 1 Project Goals (both lab and field) a)System for data acquisition, analysis, visualization and evaluation of performance parameters with cycling b)Design and validation of a mathematical model for these parameters c) Real-time and post-training analysis and visualization for bio-feedback training d) Learning of tactic approaches for a given training course profile 2 Current Progress on the Bicycle Simulator 2.1 Setup Cadence Measurement Exchangable Frame Pulse Measurement SRM Interface Simulated Gear Display Brake Unit Control Unit Elastic Suspension - Based on Cyclus 2 ergometer and our own PC-based software - Realistic pedal resistance according to GPS-based height profiles -Synchronized Video display of cycling track - Simulation of arbitrary gearshifts - Acquisition of power, cadence, heart rate - Visualization of course overviews and physiological parameters 2.2 Experiments a) Own simulator and outdoor rides, 20-25 participants - How does performance increase on specific course? - Is the simulation realistic and does it improve outdoor performance? -Which are the most effective indoor parameter displays? b) Bicycle Laboratory Freiburg - Can one conclude the pedal forces from the pedal motion? 3 Data exploration -Personal data (age, weight, size, questionnaire) -GPS, heart rate, cadence, gear, pedal force (currently 1 Hz sampling) - Motion capturing (upto 1000 Hz, 3 mm) - Lactate concentration, oxygen consumption -3D acceleration measurement to track upper body movement (approx. 50 Hz sampling) - Saddle pressure measurements - High-resolution elevation data of the Bodensee and Thurgau regions 4 Visualization Tasks existing novel on-exercise - Virtual 3D landscape or syn- chronized video playback - Superposition of values and graphics also of previous per- formances - Biofeedback information - Evaluation of efficiency of dis- plays post-exercise -Graphing values over time or distance - Few support course display as map overlay - Integration of geospatial infor- mation - Interactive visual exploration of performance data of several athletes and several training session in geospatial context - Powerwall 5 Modeling and Simulation of Adaptation Processes 5.1 Established Performance Models degra- dation recovery return to normal strain exceeded capacity Load Rate Strain Potential Response Potential Performance Potential hypertrophy limit collapse overflow + - - atrophy - Supercompensation Performance Potential 5.2 Discussion of models Feature \ Model Fit-Fat Biosynth. PerPot simulation of supercompensation stable hypertrophy overload collapse atrophy parameter access performance prediction short term long term specification for disciplines iron level, fatigue, running, tapering heart rate, hemoglobin, running, rowing, geospatial data for marathon 5.3 Ideas for advanced model a)System of differential equation (similar to PerPot but accounting for multidimensional data and including geospatial information) b) State space model (discrete hidden variables: HMM, continuous Gaussian hidden variables: Linear Dynamical System) c) Conditional Random Fields References [1]D. Saupe et al. Analysis and visualization of space-time variant parameters in endurance sports training. In Proceedings of the 6th International Symposium on Computer Science in Sports (IACSS), 2007. DFG Colloquium Thorsten Dahmen — PhD Track — Membership since 15.10.2007 Konstanz Work Group — Multimedia Signal Processing 26 June, 2008 Research Training Group 1042 (GK) — Explorative Analysis and Visualization of Large Information Spaces

Transcript of Thorsten Dahmen, Stephan Mantler · 2016. 3. 30. · Thorsten Dahmen, Stephan Mantler 1 Project...

  • Explorative Analysisand Visualization of

    Large Information Spaces

    Realistic Simulation and Optimizationof Race Bike Training

    Thorsten Dahmen, Stephan Mantler

    1 Project Goals (both lab and field)

    a) System for data acquisition, analysis, visualization and evaluation ofperformance parameters with cycling

    b) Design and validation of a mathematical model for these parametersc) Real-time and post-training analysis and visualization for

    bio-feedback trainingd) Learning of tactic approaches for a given training course profile

    2 Current Progress on the Bicycle Simulator

    2.1 Setup

    CadenceMeasurement

    ExchangableFrame

    PulseMeasurement

    SRM

    Interface

    SimulatedGear

    Display

    Brake Unit

    ControlUnit

    ElasticSuspension

    - Based on Cyclus 2 ergometer and our own PC-based software- Realistic pedal resistance according to GPS-based height profiles- Synchronized Video display of cycling track- Simulation of arbitrary gearshifts- Acquisition of power, cadence, heart rate- Visualization of course overviews and physiological parameters

    2.2 Experiments

    a) Own simulator and outdoor rides, 20-25 participants- How does performance increase on specific course?- Is the simulation realistic and does it improve outdoorperformance?

    - Which are the most effective indoor parameter displays?b) Bicycle Laboratory Freiburg

    - Can one conclude the pedal forces from the pedal motion?

    3 Data exploration

    - Personal data (age, weight, size, questionnaire)- GPS, heart rate, cadence, gear, pedal force (currently 1 Hzsampling)

    - Motion capturing (upto 1000 Hz, 3 mm)- Lactate concentration, oxygen consumption- 3D acceleration measurement to track upper body movement(approx. 50 Hz sampling)

    - Saddle pressure measurements- High-resolution elevation data of the Bodensee and Thurgau regions

    4 Visualization Tasks

    existing novel

    on-e

    xerc

    ise - Virtual 3D landscape or syn-

    chronized video playback- Superposition of values andgraphics also of previous per-formances

    - Biofeedback information- Evaluation of efficiency of dis-plays

    post

    -exe

    rcis

    e - Graphing values over time ordistance

    - Few support course displayas map overlay

    - Integration of geospatial infor-mation

    - Interactive visual exploration ofperformance data of severalathletes and several trainingsession in geospatial context

    - Powerwall

    5 Modeling and Simulation of Adaptation Processes

    5.1 Established Performance Models

    degra-dation

    recoveryreturn to normal

    strain exceededcapacity

    Load Rate

    StrainPotential

    ResponsePotential

    Performance Potential

    hypertrophylimit

    collapseoverflow +-

    -atrophy

    -

    Supercompensation Performance Potential

    5.2 Discussion of models

    Feature \ Model Fit-Fat Biosynth. PerPotsimulation of

    supercompensation 4 4 4stable hypertrophy 6 4 4overload collapse 6 4 4atrophy 6 4 4

    parameter access 4 6 4performance prediction

    short term 4 6 4long term 6 6 6

    specificationfor disciplines

    iron level,fatigue,running,tapering

    6

    heart rate,hemoglobin,running,rowing,

    geospatial data 6 6 for marathon

    5.3 Ideas for advanced model

    a) System of differential equation (similar to PerPot but accounting formultidimensional data and including geospatial information)

    b) State space model (discrete hidden variables: HMM, continuousGaussian hidden variables: Linear Dynamical System)

    c) Conditional Random Fields

    References

    [1] D. Saupe et al. Analysis and visualization of space-time variantparameters in endurance sports training. In Proceedings of the 6thInternational Symposium on Computer Science in Sports (IACSS),2007.

    DFG Colloquium Thorsten Dahmen — PhD Track — Membership since 15.10.2007Konstanz Work Group — Multimedia Signal Processing

    26 June, 2008 Research Training Group 1042 (GK) — Explorative Analysis and Visualization of Large Information Spaces