Skatelligence is an AI powered high-performance training tool for figure skaters.
Using IMUs mounted to the skater's body, Skatelligence is able to detect and categorize
the jumps they perform, and provide valuable insights into their performance.
The Skatelligence Model
6-axis IMUs mounted to the skaters torso, wrists, and skates collect and transmit live acceleration and gyroscopic
data to a local server over wifi.
The server recieves and processes the data, displaying live graphs and identifying where the
skater jumped.
Using this data, we trained a recurrent neural network to be able to classify these jumps
The neural network classfies new jumps as one of the 6 standard figure skating jumps (Salchow, Toe Loop, Loop, Flip, Lutz, Axel)
Future versions of Skatelligence will be able to identify the number of rotations
of a jump, generalize well to all skaters, and extract other useful data for training purposes