You are here

Evaluating Squat Performance with a Single Inertial Measurement Unit


Martin O'Reilly, Darragh Whelan, Charalampos Chanialidisy, Nial Friel, Eamonn Delahunt, Tomas Wardy, Brian Caulfield

Publication Type: 
Refereed Original Article
Abstract—Inertial measurement units (IMUs) may be used during exercise performance to assess form and technique. To maximise practicality and minimise cost a single-sensor system is most desirable. This study sought to investigate whether a single lumbar-worn IMU is capable of identifying seven commonly observed squatting deviations. Twenty-two volunteers (18 males, 4 females, age: 26.093.98 years, height: 1.750.14m, body mass: 75.214.2 kg) performed the squat exercise correctly and with 7 induced deviations. IMU signal features were extracted for each condition. Statistical analysis and leave one subject out classifier evaluation were used to assess the ability of a single sensor to evaluate performance. Binary level classification was able to distinguish between correct and incorrect squatting performance with a sensitivity of 64.41%, specificity of 88.01% and accuracy of 80.45%. Multi-label classification was able to distinguish between specific squat deviations with a sensitivity of 59.65%, specificity of 94.84% and accuracy of 56.55%. These results indicate that a single IMU can successfully discriminate between squatting deviations. A larger data set must be collected and more complex classification techniques developed in order to create a more robust exercise analysis IMU-based system.
Digital Object Identifer (DOI): 
Publication Status: 
Date Accepted for Publication: 
Saturday, 20 June, 2015
Publication Date: 
Wearable and Implantable Body Sensor Networks (BSN)
2015 IEEE 12th International Conference on
Research Group: 
National University of Ireland, Dublin (UCD)
Open access repository: