The aim of this study was to propose a method to automatically detect the different types of behavioural states in climbing. One climber traversed an easy route (5c difficulty on French scale) of 10 m height with a top-rope. Five inertial measurement units (IMU) (3D accelerometer, 3D gyroscope, 3D magnetometer) were attached to the pelvis, both feet and forearms to analyse the direction (3D unitary vector) of each limb and pelvis in ambient space. Segmentation and classification processes supported detection of movement and immobility phases for each IMU. Depending on whether limbs and/or pelvis were moving, four states of behaviour were detected: immobility (absence of limb and pelvis motion), hold exploration (absence of pelvis motion but at least one limb in motion), pelvis movement (pelvis in motion but absence of limb motion) and global motion (pelvis in motion and at least one limb in motion). Results indicated that the climber spent 10% of time immobile, 65% exploring holds, 1% with pelvis in motion (indicating posture regulation) and 24% in global movement (could indicate transition between holds). This new method allows automatic, rapid and reliable detection of climbing behavioural states to facilitate assessment and monitoring of climbing performance.
© Copyright 2014 Sports Technology. Taylor & Francis. Alle Rechte vorbehalten.