Athletic Performance Prediction Using Convolution Integral Scaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for predicting athlete performance are hindered by the use of arbitrary units for training intensity and duration, making it difficult to correlate these metrics with real-world performance measurements, such as power output or distance thrown.
Innovation Solution
A method and system for calculating athlete performance by receiving goal performance information, generating a training schedule, and determining a performance model based on prior performances, using training stress calculations and exponential decay constants to predict future performance, with adjustments made based on comparisons between predicted and actual performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If training stress calculations using arbitrary units (TRIMPS, TSS) are used to model athlete performance, then the system can track training volume and intensity, but the predictions cannot be directly correlated with real-world performance measurements
Solution Approach 1:
The patent transforms the arbitrary training stress units into meaningful performance predictions by changing the parameters of the convolution integral equation. Specifically, it solves for scaling factors (k1, k2) and time constants (τ1, τ2) that convert the dimensionless training stress values into actual performance metrics like power output or distance, enabling direct correlation with real-world measurements
Solution Approach 2:
The patent introduces an intermediary transformation process that bridges the gap between arbitrary training stress units and real-world performance measurements. The convolution integral serves as a mathematical mediator that processes training stress data through fitted parameters to produce performance predictions in meaningful units, rather than directly correlating the arbitrary inputs with real-world outputs
2Adaptability or versatility
If the system uses convolution integrals with multiple decaying exponential terms to model fitness and fatigue, then it can capture the complex relationship between training and performance, but the model complexity increases
Solution Approach 1:
The patent employs dynamic parameters (k1, k2, τ1, τ2) that allow the model to adapt to different athletes and training scenarios. These parameters are fitted to individual athlete data, enabling the same convolution integral framework to flexibly model various performance responses without requiring different mathematical structures for each case
Solution Approach 2:
The patent manages model complexity by fitting specific parameter values (k1, k2, τ1, τ2) to individual athlete data rather than changing the fundamental mathematical structure. This allows the versatile convolution integral model to be adapted to different athletes while maintaining a consistent, manageable framework
3Productivity
If the system predicts performance based on future training schedules, then it can provide training guidance, but the predictions require accurate knowledge of training doses that may not yet occur
Solution Approach 1:
The patent enables preliminary performance predictions by allowing the convolution integral to process planned or proposed training schedules before they are executed. The model can forecast performance outcomes based on intended training doses, providing athletes and coaches with advance guidance while maintaining reliability through subsequent validation against actual performance data
4Measurement precision
If the system transforms arbitrary performance predictions into percentile scales, then it can provide meaningful performance indicators, but the transformation process adds computational steps
Solution Approach 1:
The patent transforms arbitrary performance predictions into meaningful percentile scales by applying parameter-based scaling to the convolution integral output. The fitted parameters (k1, k2, τ1, τ2) inherently scale the predictions to match the athlete's actual performance range, and additional percentile transformations can be applied using standard statistical methods on the resulting distribution
Data Source
AI summary
A system and method of calculating athlete performance, may include receiving information relating to at least one date of performance of physical activity and generating a proposed training schedule, including one or more training sessions, corresponding to the at least one date of performance of physical activity. Further, the system and method may include receiving information relating to records of the athlete's prior performances, and determining a performance model including predicted athlete performance based on the calculated training schedule and the prior performances.