Correlation of movement patterns via pattern recognition

A structural pattern recognition method for the quantitative determination of equivalence or similarity between movement patterns was examined. A chain encoding technique was implemented for the analysis of lower limb trajectories during walking and running. Conjoint angular displacement or angular velocity patterns provided kinematic data which were cross-correlated to determine geometric congruity of within and between subject motor patterns. The correlations of the movement patterns during different speeds of locomotion revealed numerical coefficients which consistently and quantitatively discriminated the similarity or dissimilarity of limb movement patterns.