Abstract
We analyzed the skill of hula hooping motion using a Bayesian network to build a model of human periodic motion. Many human movements are periodic motions; e.g., walking and running. Therefore, an analysis method for the characterization of a periodic motion is needed. We chose the rotation of a hula hoop as periodic motion because it is easy to discriminate failure or success for this task. We created our model with the intention it be used in rehabilitation and sports instruction. Physical models are popular for motion analysis. However, they require the use of large-scale measurement systems for each application; e.g., three dimensional motion analysis, electromyograph sensing, and force-plate analysis, all require large research laboratories. Physical models are thus not always suitable. We solved this problem by building a causal relation model to focus on important body parts and important phases of movement for instruction using the Bayesian network. We found that the model can determine the skills of periodic motion.
