An Algorithm That Recognizes and Reproduces Distinct Types of Humanoid Motion Based on Periodically-Constrained Nonlinear PCA

This paper proposes a new algorithm for the automatic segmentation of motion data from a humanoid soccer playing robot that allows feedforward neural networks to generalize and reproduce various kinematic patterns, including walking, turning, and sidestepping. Data from a 20 degree-of-freedom Fujitsu HOAP-1 robot is reduced to its intrinsic dimensionality, as determined by the ISOMAP procedure, by means of nonlinear principal component analysis (NLPCA). The proposed algorithm then automatically segments motion patterns by incrementally generating periodic temporally-constrained nonlinear PCA neural networks and assigning data points to these networks in a conquer-and-divide fashion, that is, each network’s ability to learn the data influences the data’s division among the networks. The learned networks abstract five out of six types of motion without any prior information about the number or type of motion patterns. The multiple decoding subnetworks that result can serve to generate abstract actions for playing soccer and other complex tasks.