Multi-IMF Sample Entropy Features with Machine Learning for Surface Texture Recognition Based on Robot Tactile Perception

Discrimination of surface textures using tactile sensors has attracted increasing attention. Intelligent robotics with the ability to recognize and discriminate grasped objects are crucial. In this paper, a novel method for texture classification based on signals is proposed. For proposed method, first, each channel (X, Y, Z, S) decomposed empirical mode decomposition (EMD). Then, intrinsic functions (IMFs) obtained. Second, multiple IMFs, sample entropy calculated IMF. Therefore, multi-IMF (MISE) features Last but not least, two public datasets, variety machine learning algorithms used different textures. The results show that SVM MISE features, achieves highest accuracy. Undeniably, in provide idea recognition perception.