2A1-E06 ユニバーサルロボットハンドによる物体回転操作中の形状認識精度(ロボットハンドの機構と把握戦略)

概要

論文の詳細を見る
This paper describes an accuracy of shape recognition in rotating manipulation. First the time-series pressure distribution is messured with a distributed pressure sensor on the universal robot hand. Kurtosis is calculated from a pressure distribution, and the time-series kurtosis in a rotation manipulation is found. Second the time-series kurtosis is cut out with the feature kurtosis values to obtain short kurtosis patterns. The degree of similarity between these short kurtosis patterns using the DP matching recognizes the object's shape in rotating manipulation. We confirm that the recognition accuracy is high through experiments.
2007-05-11