Modelling / Simulation
Resistance training has recently become popular. If failure points, beyond which the intended motion cannot be executed, are reliably predicted, it is possible to increase the efficacy of the training and decrease the risk of injury. We aim to develop machine learning models that can enhance training effects through the proper setting of the rate of perceived exertion and prevent injuries from excessive motion by predicting the failure points. Ten young and healthy adults performed 3 sets of dumbbell arm curl using each arm with a weight of 70% of their one-repetition maximum until they reached the failure point and could not perform the standard arm curl. Using the kinematic features that we collected during each set, we developed failure prediction models based on five classification algorithms. Four models out of the five yielded the accuracy over 90%. Our findings suggest that these models can enhance the training effects by maintaining proper rate of perceived exertion, and prevent injuries due to excessive training load.
New Investigator Award
Kim, Beomdo and Ahn, Jooeun
"REAL-TIME PREDICTION OF FAILURE IN RESISTANCE TRAINING: APPLICATION TO ARM CURL,"
ISBS Proceedings Archive: Vol. 41:
1, Article 60.
Available at: https://commons.nmu.edu/isbs/vol41/iss1/60