Rollerski

Recent research has focused on using "deep" computational models to improve athlete training and performance monitoring:

: New software uses the Alpha Pose neural network to detect an athlete's motion in video footage. This "deep feature" analysis allows coaches to see precise angular movements and joint velocities that are invisible to the naked eye.

Alternatively, "Steep & Deep" is the name of a specific type of Cross Roller Ski Race that takes athletes off smooth tarmac and onto more challenging, irregular terrain. These events often emphasize:

: Deep machine learning is also applied to classify specific skiing gears (like G2 or G3 techniques) during uphill climbs based on sensor data. "Steep & Deep" Race Events

: Using "all-terrain" rollerskis like the NORDICX Hybrid Skate which feature larger pneumatic wheels (125mm–200mm) for gravel or rough paths . Performance & Safety Features

: Researchers use Deep Long Short-Term Memory (LSTM) neural networks to estimate an athlete's power output. This is done by processing data from wearable Inertial Measurement Units (IMUs) worn on the body.

: Taking advantage of iconic, steep ascents that aren't accessible for skiing in winter.