Rider motion identification during normal bicycling by means of principal component analysis

Recent observations of a bicyclist riding through town and on a treadmill show that the rider uses the upper body very little when performing normal maneuvers and that the bicyclist may, in fact, primarily use steering input for control. The observations also revealed that other motions such as lateral movement of the knees were used in low speed stabilization. In order to validate the hypothesis that there is little upper body motion during casual cycling, an in-depth motion capture analysis was performed on the bicycle and rider system. We used motion capture technology to record the motion of three similar young adult male riders riding two different city bicycles on a treadmill. Each rider rode each bicycle while performing stability trials at speeds ranging from 2 km/h to 30 km/h: stabilizing while pedaling normally, stabilizing without pedaling, line tracking while pedaling, and stabilizing with no-hands. These tasks were chosen with the intent of examining differences in the kinematics at various speeds, the effects of pedaling on the system, upper body control motions and the differences in tracking and stabilization. Principal component analysis was used to transform the data into a manageable set organized by the variance associated with the principal components. In this paper, these principal components were used to characterize various distinct kinematic motions that occur during stabilization with and without pedaling. These motions were grouped on the basis of correlation and conclusions were drawn about which motions are candidates for stabilization-related control actions.