10th EAI International Conference on Pervasive Computing Technologies for Healthcare

Research Article

Evaluating Performance of the Lunge Exercise with Multiple and Individual Inertial Measurement Units

  • @INPROCEEDINGS{10.4108/eai.16-5-2016.2263319,
        author={Darragh Whelan and Martin O'Reilly and Tom\^{a}s Ward and Eamonn Delahunt and Brian Caulfield},
        title={Evaluating Performance of the Lunge Exercise with Multiple and Individual Inertial Measurement Units},
        proceedings={10th EAI International Conference on Pervasive Computing Technologies for Healthcare},
        publisher={ACM},
        proceedings_a={PERVASIVEHEALTH},
        year={2016},
        month={6},
        keywords={exercise classification inertial measurement units lunge functional screening tools biofeedback rehabilitation},
        doi={10.4108/eai.16-5-2016.2263319}
    }
    
  • Darragh Whelan
    Martin O'Reilly
    Tomás Ward
    Eamonn Delahunt
    Brian Caulfield
    Year: 2016
    Evaluating Performance of the Lunge Exercise with Multiple and Individual Inertial Measurement Units
    PERVASIVEHEALTH
    EAI
    DOI: 10.4108/eai.16-5-2016.2263319
Darragh Whelan1,*, Martin O'Reilly1, Tomás Ward2, Eamonn Delahunt3, Brian Caulfield1
  • 1: Insight Centre for Data Analytics, University College Dublin
  • 2: Insight Centre for Data Analytics, National University Of Ireland, Maynooth
  • 3: School of Public Health, Physiotherapy and Sports Science, University College Dublin
*Contact email: darragh.whelan@insight-centre.org

Abstract

The lunge is an important component of lower limb rehabilitation, strengthening and injury risk screening. Completing the movement incorrectly alters muscle activation and increases stress on knee, hip and ankle joints. This study sought to investigate whether IMUs are capable of discriminating between correct and incorrect performance of the lunge. Eighty volunteers (57 males, 23 females, age: 24.68 +/- 4.91 years, height: 1.75 +/- 0.094m, body mass: 76.01 +/- 13.29kg) were fitted with five IMUs positioned on the lumbar spine, thighs and shanks. They then performed the lunge exercise with correct form and 11 specific deviations from acceptable form. Features were extracted from the labelled sensor data and used to train and evaluate random-forests classifiers. The system achieved 83% accuracy, 62% sensitivity and 90% specificity in binary classification with a single sensor placed on the right thigh and 90% accuracy, 80% sensitivity and 92% specificity using five IMUs. This multi-sensor set up can detect specific deviations with 70% accuracy. These results indicate that a single IMU has the potential to differentiate between correct and incorrect lunge form and using multiple IMUs adds the possibility of identifying specific deviations a user is making when completing the lunge.