Header Ads

Cycling Training Software: Scientific Research and Experimental Development Project Description (2006)

To develop training methods and devices for competitive cyclists based on improved data collection and tracking systems, incorporation of improved performance monitoring devices and integration of scientific training tools for endurance sports.

A. Scientific or Technological Objectives
To develop training methods and devices for competitive cyclists based on improved data collection and tracking systems, incorporation of improved performance monitoring devices and integration of scientific training tools for endurance sports.

B. Technology or Knowledge Base Level
Available performance monitoring systems are merely measurement tools and do not provide any detailed data analysis or criteria to evaluate athletic performance. Measures of athletic performance, such as heart rate or average velocity, may be inadequate when taken singly. However if such measures can be combined with power and cadence in a single data collection and analysis platform, it may be possible to develop advanced training routines for improved performance.
Heart rate monitors have been used since the 1980s to improve training performance and provide feedback.  While heart-rate is a good indication of intensity, an athlete’s heart-rate is affected by many uncontrolled variables.  These variables include: temperature, hydration, fatigue, illness, lack of sleep, food intake, jet-lag and altitude.  When used in conjunction with power output, heart-rate can be very useful in determining output and the state of the athlete.
Power is a direct reflection of the exercise intensity.  From power calculations, an accurate picture of how efficiently the body is performing can be obtained.  On this basis, whether racing or training, one exercise session can be compared to another.  Existing software for power meter systems allow for data analysis, but are not incorporated into a training program.  An experienced coach is required to analyse the data and recommend adjustments to the training program.  The training software is independent of the power meter software.
Another useful tool in determining intensity is blood lactate levels.  These are typically taken during lab tests or at the end of a hard effort.  Portable lactate testers can measure the amount of lactate acid in the blood, providing an accurate indication of the intensity of the effort.  These measurements can be done by a coach, but are harder to do as extreme care must be taken by the coach to avoid contamination by the athlete’s blood and ensuring that the athlete does not get infection.  The lactate levels can be compared to heart-rate and power out-put.
Several training tools have been shown in the scientific literature to be valid and reliable, however they have not been used in training software packages. When integrated with the algorithms developed in this project, these tools will provide coaches and scientists with the ability to perform valid scientific research on the effect of training on performance.
The first scientific tool is the Borg scale of Perceived Exertion. It is a 0-10 scale that athletes can use to rate how hard an effort felt. It is the only scale that allows a statistical analysis to be done on the results, and as such it is the gold standard within sports science. By using this scale rather than an arbitrary one, this software tool will become the first training software to provide the ability for scientists to use it for research purposes, taking the field of Kinesiology forward.
The second scientific tool is the Daily Analysis of Life Demands of Athletes (DALDA) questionnaire. This has been validated for use with athletes, including Olympians. It assesses both the sources of stress and the symptoms of stress felt by an athlete. It uses a range of “normal” variation that is individual to each athlete to determine how that person is tolerating a given training load. It has been used to identify early signs of overtraining, to identify signs that an athlete isn’t training hard enough, and to monitor periods when an athlete is undertaking a taper leading into competition. The goal of this questionnaire is to optimize both health and fitness in athletes.
The integration of the Borg Scale and the DALDA questionnaire into the algorithms developed in this project will represent a major advance in the field of Kinesiology. It has not been possible in the past to concurrently measure physiological data (from power meters or heart rate monitors) together with the athlete’s perception of effort and stress. The heart of training science is the interaction of these two aspects, and this software will make advanced, rigorous study of this interaction possible.
C. Scientific or Technological Advancement
The following advancements are sought in the field of sports science:
1. To develop a comprehensive post-processing software suite that will allow for each cyclist’s performance to be analysed and compared over multiple training sessions both individually and as compared to multiple athletes.
Performance outputs will vary from day to day. A “bad day” may be the result of high wind velocity, colder temperature, extreme heat, poor equipment choices, inadequate nutrition, poor road surface, or poor tactics.  The body will also have a lower output with lack of sleep, stress, extreme distractions or fighting an illness (not noticeable at times because of a lack of symptoms).  Lower out-put needs to be addressed or monitored immediately to avoid a long-term rest period with no training.
Algorithms will be developed to differentiate between low performance due to one bad day or due to lower overall fitness.  Over time an athlete should have a more consistently higher output.  With existing data analysis, it is hard to differentiate a low fluctuation to an over-all lower performance.  Low performance trend must be rapidly identified such that changes can be made to training.
Current performance indicators do not provide sufficient warning to enable rapid changes in the training load before excessive fatigue occurs.  Once an athlete is over-trained they require serious measures to recover including long periods of complete rest.
The system must monitor all factors which can be controlled by the athlete. This includes monitoring the athlete’s fatigue level by adjusting training load and recovery in a dynamic process. Monitoring each training session and having software that detects abnormalities right away will allow timely adjustments to be made.
2.   To develop methods of using collected data to accurately assess power expenditure and allow for efficiency analysis for individual athletes.
Human bodies react differently to outside temperature, rain, wind, and altitude.  Identifying how an individual athlete reacts will allow focused training to maximize performance at an event which may include a weakness.  Collecting training data over time will show factors which may affect an individual athlete.  Analyzing data based on environmental conditions will allow an athlete to prepare for all competition situations and variables.  Training algorithms must be developed to evaluate power outputs/heart-rate in different environmental conditions.
Sensor data will be evaluated to determine features of motion and physical performance. Existing power analysis software includes a way to analyse an athlete’s pedaling efficiency by evaluating the power input from both legs.  Currently this can only be analyzed in a specific mode and is not part of a regular evaluation.  While high level cyclists tend to have a similar power out-put from both legs, this balanced out-put takes a long time to develop.  The system will evaluate pedal efficiency data as part of regular daily data collection in order to develop training methods to achieve balanced output.
3.   To develop detailed training regimes for cyclists using the performance indicators derived from the post processing software.  The key to this training program will be the dynamic evaluation of the power/heart-rate, the integration of advanced tools and modification of the training program to achieve the desired short and long-term results. Investigation will be conducted to determine relationships between physiological (from power meters or heart rate monitors) and an athlete’s perception of effort and stress.
Data will be aggregated from all athletes monitored in order to determine any patterns that are indicative of a high performing athlete. The system will readily enable data characteristics and trend common to all successful athletes to be identified. Athletes demonstrating the desired performances will be compared using specific parameters, in order to find keys to focus on in the developing athletes.
D. Work Performed in this 12-Month Period
The following work was conducted in this period.
Work was conducted on the development of a post-processing software suite that would allow for each cyclist’s performance to be analysed and compared over multiple training sessions both individually and as compared to multiple athletes.
Existing analysis software only compared individual work-outs.  They could be compared manually by switching between different work-outs.  Algorithms were required to automatically track work-outs over a period of time thus speeding up data comparison.  Algorithms were sought to track multiple athletes in many different areas by permitting a trainer to specify the factor or factors to be compared.
A system was developed that could also alert a coach to variations of an athlete’s output trend.  This could indicate an improvement or decrease in fitness.  The training load of an athlete needs to be adjusted with adaptations to have continued improvement.  It is critical to adjust training loads if an athlete is fighting fatigue or illness to avoid overtraining. Algorithms that can indicate changes in an athlete quickly will result in a training program which will maximize the athlete’s time and performance.
The development of the algorithms was done in conjunction with lab ramp tests and road tests. During lab tests, an athlete on a stationary bike was monitored while undergoing testing at a variety of intensities. Both the athlete and the stationary bike were fitted with sensors to monitor performance. During the road tests, the athlete was followed by a van in order to collect data from the sensors during training exercises. Data from these sensors was analysed by the software developers and Kinesiologists to modify the algorithms and make changes to the training regimens.
A number of sensors from various manufacturers were mounted various bicycles in order to resolve uncertainties relating to the integration of sensor data from power meters and heart rate monitors into the code. These sensors included Polar bicycle sensors, Power Tap recording devices, SRM recording devices and other heart rate/ power meter devices. A PDA device was used to synchronize devices. These sensors required proper installation and maintenance for accurate operation.  If the metres are not calibrated as per the manufacturer’s protocol, they will result in inaccurate results.  The power meters also occasionally malfunction.  The power out-put will either show discrepancies too-high/too-low if they are improperly calibrated. Algorithms were developed to monitor the power meters to enable irregularities to be identified.
Detailed training regimes were developed for cyclists using the performance indicators derived from the post processing software.  The process for developing training regimes was as follows:
  1. Ramp tests were conducted athletes on a test bike in a laboratory.  This involved using a measured power load (watts) and increasing the load to a maximum effort.  Heart-rate was measured during the test and blood-lactate was tested at the end of each ramp.  This test determined the athletes MAP (maximum aerobic power), lactate threshold and heart-rate at the different loads.
  2. The athlete performed a time-trial on a set course, using a power meter to determine power-output and average heart-rate.
  3. Training zones were determined with the help of a physiologist.  These zones were defined in terms of watts (power out-put) and heart-rate.
  4. Based on the tests and the desired performance goals set by the athlete and coach, a training program was created.  The specific work-outs could be determined for four week periods at a time.  A yearly plan could be made to break-up preparation into different phases based on a competition schedule and the athlete’s performance goals.
  5. Daily training data including data from the training session (power-meter, heart-rate), morning rest, perceived efforts, time of training, time of day, and environmental conditions were automatically analysed by the algorithms, compared and provided to the coach.
  6. Based on this data, modifications to the training program could be made to account for performance improvements or deviations from the training program.  Algorithms were built to track the trends in power out-puts for work-outs and show the changes over any desired time period.
  7. Tests were performed approximately every 2 months in the lab to track each athlete’s progress.
  8. The key to the training program was to be the dynamic evaluation of the power/heart-rate and modification of the training program to achieve the desired short and long-term results.
Methods were developed of using collected data to accurately assess power expenditure and allow for efficiency analysis for individual athletes. By incorporating environmental factors into the development, algorithms were developed to evaluate power outputs/heart-rate in different environmental conditions.
Sensor data was evaluated to determine features of motion and physical performance. Existing power analysis software includes a way to analyse an athlete’s pedaling efficiency by evaluating the power input from both legs.  Currently this can only be analyzed in a specific mode and is not part of a regular evaluation.  While high level cyclists tend to have a similar power out-put from both legs, this balanced out-put takes a long time to develop.  Algorithms were created to evaluate pedal efficiency data as part of regular daily data collection in order to develop training methods to achieve balanced output.
Data was aggregated from all athletes monitored in order to determine any patterns that are indicative of a high performing athlete. The system will readily enable data characteristics and trend common to all successful athletes to be identified. Athletes demonstrating the desired performances will be compared using specific parameters, in order to find keys to focus on in the developing athletes.
Data was analysed in order to provide feedback as to whether an athlete’s performance level is increasing or decreasing.  Comparing power out-put’s relationship to heart-rate was of critical importance.  For example: An athlete may typically have a heart-rate of 167 bpm @ 300 watts (power output).  If the athlete is fatigued the heart-rate may be at 175 bpm with the same power out-put.  This shows that there is an abnormality with the athlete, which needs to be addressed.  Over time the same athlete should have a lower heart-rate with the same power out-put which indicates improved fitness.  At this time this data can be compared manually and is quite labour intensive.
Training an individual athlete requires a delicate balance of training loads which may cause short-term fatigue and excessive loads, which may cause a state of over-training.  With proper recovery, the training load pushes the athlete to adapt to a higher level of fitness.  Over-training occurs when an athlete does not get enough recovery from hard efforts.  Over time, the body tries to adapt, but starts to deteriorate.  This can be a slow process and once it is detected may require two weeks to two months of total recovery.  The system must enable close monitoring to prevent over-training.
Power/Heart-rate comparison:
  • Shows if training zones are relevant
  • Indicates fatigue on a short-term basis
  • Shows if an athlete is ready for hard training or requires rest
  • Indicates improvements in efficiency and increased MAP/VO2Max
  • Is critical to dynamic coaching adjustments
  • Based on base-line information and on percentage of max heart-rate
  • Shows improved/decreased fitness over a period of time
  • Helps determine an over-training state at an early stage
Algorithms were developed to differentiate between an individual day with lower performance and a low performance trend taking into account a number of factors.
At the end of the fiscal period, work on developing and refining the algorithms was continuing. Ongoing experimentation will involve the integration of scientific monitoring tools into the training logging and planning algorithms. The integration of the Borg Scale and the DALDA questionnaire into the software tool will represent a major advance in the field of Kinesiology. It has not been possible in the past to concurrently measure physiological data (from power meters or heart rate monitors) together with the athlete’s perception of effort and stress. The heart of training science is the interaction of these two aspects, and this software will make advanced, rigorous study of this interaction possible. When integrated with this program, these tools will provide coaches and scientists with the ability to perform valid scientific research on the effect of training on performance.
At the time of this writing, this software tool has now provided a breakthrough in the integration of training data analysis and monitoring with training planning. It provides an easy interface between athlete and coach, removing the obstacles often encountered when athletes train and/or compete away from their coach. This has allowed coaches to review and adjust their training load much faster than in the past.
These new advances in the software tool will allow athletes not merely to plan and log training, but to monitor, review, and learn from it in a comprehensive way that is currently not available on the market.  This project will result in a major breakthrough for athletes training in endurance sports such as running, swimming, cycling, rowing, weight lifting, hiking, cross-country skiing and speed skating.
E. Supporting Information
Retain all project information including invoices, technical reports, test data and project timelines.


   
Media

  


  


 


   




No comments

Powered by Blogger.