Tactile texture recognition with a 3-axial force MEMS integrated artificial finger

Abstract

Abstract-Recently, several three-axial MEMS-based force sensors have been developed. This kind of force micro sensor is also called tactile sensor in literature for its similarities in size and sensitivity with human mechanoreceptors. Therefore, we believe these three-axial force sensors being able to analyse textures properties while sliding on a surface, as would do a person with his finger. In this paper, we present one of these sensors packaged as an artificial finger, with a hard structure for the bone and a soft rubber for the skin. Preliminary experiments show a good sensitivity of the finger, as its ability to sense the periodic structure of fabrics or to differentiate papers from fabrics calculating a friction coefficient. Its performance for discrimination of different surfaces is then estimated on fine textures of 10 kinds of paper. Supervised classification methods are tested on the data. They lead to an automatic classifier of the 10 papers showing good performances.

Tactile object class and internal state recognition for mobile manipulation

“…Others have used tactile sensors to detect ridges and bumps in the material [16] by sliding the robotic finger over an object to estimate the material from the obtained frequency spectrum [4]. Sensors based on piezoelectric vibration have been used to determine the hardness or softness of probed biomaterials [17].…”

Face classification using touch with a humanoid robot hand

“…In order to do so, our humanoid robot -the iCub robot [12] -moves its hand in a vertical plane, while its fingers actively follow the curve of the face to track its shape 1 .There are four faces to classify in this experiment (see Figure 1).…”

“…Recent progresses in tactile sensing offer a range of novel research directions in robotics, such as recognition of objects through texture [1] and safety during physical human-robot interaction [2]. In humanoid robots, the sense of touch becomes even more important due to the resemblance of these robots to humans and the fact that touch is an essential sense for humans.…”

Grasp stability assessment through unsupervised feature learning of tactile images

“…For instance, researchers have used tactile sensors to recognize object textures [8], and to detect object slippage [9], [10], [11]. Another component would be the ability to evaluate grasp stability, and use this knowledge to predict grasp failure.…”