Abstract
Coordination of the whole body movement is important for improvement of front crawl performance. The arm pull, body rolling and flutter kick are characteristic motions in the coordination of the front crawl. In addition, it is difficult to master the appropriate kick timing because the rhythm of flutter kick is quicker than the other motions. However, the training to achieve the appropriate coordination is generally conducted empirically. Therefore, it is useful for swimmers to develop a bio-feedback training system which indicates the appropriate leg kick timing estimated from the rolling motion. In this report, the algorithm which estimates kick timing from roll motion using a CPG network with neural oscillators were constructed. The roll and hip joint angles could be generated by the proposed CPG network. In addition, roll angular velocity during six beat front crawl measured in an experiment were put into the CPG network. The CPG network could generate the roll and hip joint angles synchronized with the swimmer's roll motion.
