触覚情報を利用した粉粒体の粒径識別 : ヒト触覚受容器の機能を考慮した人工システムの粒径推定アルゴリズム(機械要素,潤滑,工作,生産管理など)

概要

論文の詳細を見る
Discrimination of particles has been investigated by using tactile information in an artificial system. Two kinds of time-series signals, i.e. normal and tangential fluctuations, were obtained experimentally when two silicone-gum specimens were rubbed with sands between the surfaces of specimens, in which the sands with sizes of 20-710μm were classified into nine samples by sieves. The particle diameters of samples were found to be correlated with two quantities obtained from the time-series signals; one was the normal elastic recovery between before and after rubbing, and the other was the weighted mean frequency of dominant elements appearing in the spectrum of tangential fluctuation. By using the relationship between the two quantities and the particle diameters, the artificial system showed an ability of discrimination comparable with human subjects. Information of the size of sands appears in three ways. In case of larger-size samples, it appears strongly in the weighted mean frequency with lower-range frequency. On the contrary, it appears strongly in the elastic recovery for middle-size samples, and in the weighted mean frequency with higher-range frequency for smaller-size samples. By considering such characteristics, an algorithm has been constructed for discriminating the all samples with a higher accuracy, which has a similarity to the functions of three kinds of tactile receptors.
社団法人日本機械学会の論文
2007-01-25