|
|  |
Powerbike
Acquisition, modeling, and visualization of
performance parameters in race-bike training
Work Group Multimedia Signal Processing
Prof. Dietmar Saupe
Dipl.-Ing Thorsten Dahmen
M.Sc. Roman Byshko
Cooperations:
Prof. Hartmut Riehle (Sports Science)
Dr. Dietmar Lüchtenberg (Sports Science)
Christian Federolf (Sports Science)
Niki Cerha (Sports Science)
Prof. Daniel Keim (Computer Science), Powerwall Installation
Summary
We develop methods for data acquisition, analysis, modeling and visualization of performance parameters in endurance sports with emphasis on competitive cycling. The goal of our work is to provide methods to visualize, analyze, model, and improve the performance of athletes in training and in competition. Measurements from a palette of devices, including common bike computers, GPS-recorders, power meters, and spiroergometric devices are combined requiring data fusion and synchronization.
In addition, we consider integration of maps with elevation information and areal photographs together with recorded video footage of training courses. We also apply biofeedback methods that require complex data processing and real-time visualization. Appropriate ways to jointly present such diverse data need to be developed for their analysis and visualization. A particular challenge is the information visualization of complex multi-dimensional signals of traveling sensors. For the simultaneous visualization of large information quantities we use the Powerwall at the University of Konstanz, a large high resolution display, which offers the display of high resolution terrain data and maps together with static and dynamic parameter sequences from measured training rides or a running biofeedback simulation.
The following topics are addressed among others:
- development of a multimodal and scalable system for acquisition of performance parameters, their visualization and analysis, both in the lab and in the field
- designing, validating, and applying a mathematical model for these performance parameters
- adaptation of training course profile for laboratory biofeedback simulation
- learning of tactic approaches for a given training course profile
- evaluation of efficiency of different visual feedback parameters
Motivation
Computer science in sports is an emerging interdisciplinary field which has evolved in the last 20 to 30 years focusing on the following areas of research: data acquisition, processing and analysis, modeling and simulation, data bases and expert systems, multimedia and presentation, and IT networks/communication. Recording devices for a host of physical and physiological parameters have become available both to the professional athlete as well as to hobby and amateur sportsmen. These are being used for monitoring and measuring sports activities in the lab, during training and even in competitions.
But after the data are available it still remains difficult to extract the important or relevant information out of it. In this project we contribute to the research aimed at the entire cycle of data acquisition, filtering, analysis, visualization, modeling, and prediction of such complex data. We selected endurance sports as a particularly suitable type of sport because of its character allows for long time series of data that are expected to be more homogeneous and depend to a lesser degree on chance events as, e.g., in game sports. Our project starts out focusing on road cycling and may be extended later to also cover competitive running and rowing.
Current Progress
We developed a bicycle simulator based on a Cyclus 2 ergometer
and our own PC-based control software. The main components of the simulation are
a computer controlled pedal resistance according to the height profile of a
cycling track, and a video display of the cycling track that shows the current
position. The main motivation behind the development of the simulator are to
allow the measurement of training parameters in a laboratory environment,
to familiarize cyclists with unknown cycling tracks to an extent that allows
competitive performance on that track, and to develop models for training
control and performance prediction.
We designed this software for platform independence
and display scalability. It runs on several operating systems (Linux, Mac OS,
Windows), on normal PC screens, and on the Konstanz Powerwall. The Konstanz
Powerwall is a large back projection screen lit by eight self-aligned beamers.
The total resolution of the Konstanz Powerwall is 4640 x 1920 pixels.
The main functions of the simulator software are the control of the
pedal resistance of the ergometer, the video playback synchronized to the track
position, the recording of training data measurements (speed, cadence,
power, heart rate, etc.), and visualization of recorded training data.
For the simulation of a course, we use data from a number of different sources.
The recording of a GPS device gives us the height profile of the course as well
as the exact geographic position. The geographic position is used to extract a
digital terrain model of the course from a geographic information system, such as Google Maps,
from which we render a 3D-overview of the course. A third data source is a video
recording of the course, which is captured at the same time as the GPS recording and
is later synchronized to it with the help of the recording time stamps in the data.
The data processing of the simulator was also the topic of a Bachelor thesis in our
group.
An early version of the simulator software was presented
at the 6th Intern. Symposium Computer Science in Sport (IACSS'07, Calgary, June 3--6, 2007).
A functional version of the simulator was presented at the opening of the new
fair grounds in Stuttgart, October 19--21, 2007.
In 2007, we participated in a yearly training seminar held by the sports department
of the University of Konstanz in the Engadin valley (Swiss Alps). During the seminar,
we collected data of two major alpine passes (Flueelapass and Albulapass)
neighboring the Engadin valley. The data collection was done by GPS and video.
We used a GPS device with recording capability (Garmin Edge 305) to record the
geographic position and the height profile of a track and a video camera to capture
the view of a cyclist on the track. Later, we also collected data of three popular
tracks in the surrounding area of Konstanz (Ottenberg, Schiener Berg,
and Allensbach-Langenrain). Another part of the seminar were spiroergometric measurements.
In a cooperation, Dr. Nikolaus Seltzer from the Herzzentrum Lahr conducted a series
of spiroergometric measurements at increasing elevations above sea level.
We use the tracks in the surrounding of Konstanz to validate the accuracy of the
simulation in terms of pedaling power. We compare the total pedaling energy required to
climb the Ottenberg on a real bicycle and to complete the Ottenberg track on the
simulator. The energy is measured with an SRM power meter that is mounted to
the bicycle used on the real Ottenberg and later to the ergometer of the bicycle
simulator. The SRM power meter measures the pedaling power from which the total
pedaling energy can be calculated by integration over the measured time series.
So far, we did a validation run on the Ottenberg course, with accuracy results in the
range of the accuracy of the SRM power meter itself.
Currently we take measurements from a group of sports science students who ride the Schiener Berg both on our simulator and in reality several times under constant conditions. Simultaneously we develop new software in order to chart and analyze the data. The aim is to answer the following questions
- How does performance increase on specific course?
- Is the simulation realistic and does it improve the outdoor performance?
- Do the displayed indoor parameters help the cyclist?
- Which displays are most effective?
Goals
In professional training for competitive cycling measurements are taken that capture physical parameters of the athlete and mechanical parameters of the bike in the field or of an ergometer in a laboratory environment. Such data series are usually transferred to a computer for display and analysis by special purpose software. Systems focus on monitoring and analyzing those performance parameters that are characteristic for the specific motions in biking.
The goal is an improvement of cycling performance in training and eventually in competition. There is a rapidly growing palette of commercial measuring devices, ranging from common bike computers capturing speed, cadence, heart rate, temperature, and barometric pressure to more complex ones for GPS localization. Moreover, there are (expensive) power meters and spiroergometric devices that measure ventilation and gas exchange.
For scientific research it is desirable to combine several such devices, thus, creating the need for methods for data fusion and synchronization. Commercial metering devices come with specific software that does not readily support the desired joint representation and analysis. In our work we strive to combine all types of data with high-resolution maps of training regions and corresponding video recordings synchronized by software to individual training rides. The amount of data collected can get very large, will have various dimensionality (time series, geospatial data, video, etc.), must be processed in real-time for use during training and will go through elaborate cycles of visualization, analysis, and modeling, has varying precision and may be contaminated by noise and may be incomplete. For the simultaneous visualization of large information quantities we apply the Powerwall at the University of Konstanz, i.e., a large high resolution display (dimensions 5.20 m x 2.15 m, pixel resolution 4640 x 1920) which offers static or dynamic display of high resolution terrain data and maps together with parameter sequences from measured training rides.
On the other hand, online visualization and operation in the field must be restricted to very small battery powered displays. All these constraints imply that highly scalable methods for appropriate analysis and visualization are required. From the computer science side the challenges are:
- to custom tailor tools for the acquisition, synchronization, fusion, and management of large numbers of time series
- to devise scalable solution strategies for the visual combination of terrain data, video, simultaneous time sequences of many parameters on the Powerwall and on regular size computer screens, ensuring to communicate the massive data appropriately for the different types of users (athletes, sports scientist, trainers, general public)
- to develop models for performance in road cycling that can be used to plan training sessions and predict performance in the field, as well as to redesign and improve the visualization system
- to develop real-time analysis and visualization techniques including appropriate human-computer interfaces tailored for application in the field (on a programmable handle-bar mounted system) and in the lab for biofeedback training sessions
- to organize the visualization of past training sessions and online biofeedback training sessions suitable for the domain experts, i.e., for training science, resp. the athlete
Publications
M. Fueller. Datenfusion und Synchronisation. Bachelor's Thesis, University of Konstanz, 2006
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), 2007.
|
