Analysis of touching motion using singular spectrum transformation

Recently, robots have been increasing demand for elderly care problem, dangerous work involved in life, disappearing skilled work and so on. To use robots like human, human motion information is needed to obtain. The motion information is also needed to analyze in order to use effectively information. In this paper, motion information is extracted by motion-copying system. Motion-copying system is suitable for obtained touching motion. To analyze motion information, Singular Spectrum Transformation (SST) is introduced. SST is one of the most adaptive data mining techniques to analyze time-series data. Motion information has various parameters in time-series data. This mining system has probability to use for motion segmentation and classification. According to the experiments and the analysis, touching motion can be analyzed.