|
Using global positioning systems for modelling athletic performances| old_uid | 19757 |
|---|
| title | Using global positioning systems for modelling athletic performances |
|---|
| start_date | 2021/11/25 |
|---|
| schedule | 14h |
|---|
| online | no |
|---|
| summary | Wearable global positioning systems (GPS) are gold standard systems of measurements in outdoor team sports. Tracking the position of a player with precision allows for monitoring athletes in several ways. To date, monitoring from GPS mainly relies on summarised statistics of sessions based on expert empirical knowledge. Predictive applications using GPS data remain, however, new and focused on injury predictions.
In this seminar, we show how athletic performance could be extracted and predicted directly from GPS measurements.
From individual acceleration - velocity profiles modelled from games data, we processed three prediction tasks with (i) times-series forecasting from games data, (ii) regression and classification tasks from commercial summarised statistics, and (iii) a regression task using features extracted from raw GPS data.
We will discuss the results in light of the multi-factorial aspects of athletic performance and address future lines of researches. |
|---|
| responsibles | Marin |
|---|
| |
|