Van Anh Ho

Development and Analysis of a Sliding Tactile Soft Fingertip Embedded With a Microforce/Moment Sensor

We describe the development of a tactile hemispherical soft fingertip (FT) of a size similar to that of a human thumb. The sensory core consists of a microscaled force/torque sensor that can output one component of force and two components of moment simultaneously, which was developed beforehand. This sensor is embedded in a polyurethane rubber hemispherical dome to form a complete soft, compliant, and perceptible robotic FT. This system is designed for easy fabrication, high reliability in outputting signals, and stable operation. Static and dynamic mathematical analyses were utilized to investigate the responses of the sensor during the typical sliding motion of an FT. This was followed by experiments to show its potential in tactile and texture recognition. Especially, incipient-slip detection, which is critical in grasping manipulations, can be assessed properly and in a timely way. The development of this tactile FT is considered significant in the field of dexterous manipulation.

Investigation of a biomimetic fingertip's ability to discriminate fabrics based on surface textures

Tactile sensing is an important ability for a humanoid robot which interacts in an unstructured environment. Such a system needs to sense and evaluate surface properties of objects that it interacts with. Among those properties, surface texture identification is a compulsory ability for certain kind of robot systems such as service robots, medical robots and exploratory robots. Therefore, the tactile system of above type of robots should have the ability to identify and discriminate textures with acceptable accuracy. A biomimetic fingertip that can be used in above kinds of robot tactile systems is introduced. The fingertip has the ability to detect force and vibration modalities. This paper reports the ability of the fingertip system to discriminate multiple materials (six fabrics and aluminium plate) by comparing the differences in their surface texture. The materials were classified using features: variance and power of the accelerometer signal. Moreover an Artificial Neural Network (ANN) classifier was evaluated by using the first 300 Fourier coefficients of the accelerometer signal as features. Above two methods used the raw signals of the accelerometers nearest to the contact area. Finally, an input signal was computed by calculating the covariance of two adjacent accelerometers. This new signal was used to calculate features for the ANN classifier. The results showed that the use of a convoluted signal improved the success rate of discriminating the seven textures.

What can be inferred from a tactile arrayed sensor in autonomous in-hand manipulation?

In this paper, we present a concept of using tactile sensors as sufficient tools in localizing, recognizing an object in robotic in-hand manipulation tasks. Our approach operates on a moderately intensive array data that are obtained from a tactile sensor when a robotic gripper grasps an object that is small relatively to fingers. In stead of using tactile data as an array of discrete numbers, we treat it as a grayscale image. By working with successive images from tactile sensor exploiting image processing tools, we are able to extract rich information about the contact condition between an object and the gripper. Experimental results show that from the processed data, we can realize the grasped objects position/orientation, contact shape, especially the stick-slip condition on the contact surface that is derived for the first time for this sensor. Based on localized displacement phenomenon of a sliding soft object, we proposed an approach to detect the slippage quantitatively, and conducted a wide range of experiment in term of characteristics of objects movement. We also conducted a model for an object-grasping gripper with tactile feedback in various postures of the object, and a corresponding experiment setup to validate computed results. Presented result is expected to be used widely in the community due to its simplicity and reliability.

A 3-D Nonhomogeneous FE Model of Human Fingertip Based on MRI Measurements

A 3-D nonhomogeneous finite-element (FE) dynamic model of a primate fingertip is developed in this paper based on magnetic resonance (MR) imaging measurements for better understanding the mechanism of human finger sensation. The geometries of a human fingertip are measured using an MR system, and a series of 2-D images is obtained. Utilizing a boundary tracking method, boundaries of the fingertip and distal phalanx are tracked from each image slice and a set of boundary nodes is generated to construct a 3-D tetrahedral mesh of the fingertip. The 3-D mesh is then utilized to formulate a nonhomogeneous FE dynamic model for simulating the fingertip behaviors. The constitutive model, which consists of elastic and viscous elements, is employed to govern the dynamic behaviors of individual tetrahedral FE. The FE model is further extended to deal with contact interaction between the fingertip and an external instrument. Differing with conventional fingertip models, the model presented in this paper is able to not only better represent internal and external geometries of a human fingertip but also take the tissue viscosity into consideration. Simulation and experiments are performed with both a human finger and a fingertip phantom under different indentation operations. We found that the Voigt model can simulate the force behaviors of a fingertip phantom but has difficulty to reproduce the force relaxation behavior of a human fingertip. We have therefore introduced a parallel five-parameter physical model to solve this problem.

Flexible Fabric Sensor Toward a Humanoid Robot's Skin: Fabrication, Characterization, and Perceptions

We have developed a fabric sensor knitted of tension-sensitive electro-conductive yarns. Each yarn has an elastic core, around which is wound two other separate, tension-sensitive electro-conductive threads, making this sensor inherently flexible and stretchable and allowing it to conform to any complicated surface on a robot, acting as a robotic skin. The pile-shaped surface of the sensor enhances its ability to detect tangential traction, while also enabling it to sense a normal load. Our aim is to use this sensor in applications involving relative sliding between its surface and a touched object, such as contact recognition, slip detection, and surface identification through a sliding motion. We carefully analyzed the static and dynamic characteristics of this sensor while varying the load and stretching force to fully understand its response and determine its degree of flexibility and stretchability. We found that a discrete wavelet transformation may be used to indicate stick/slip states while the sensor is sliding over surfaces. This method was then used to detect slippage events acting on the sensors surface, and to decode textures in a classification test using an artificial neural network. Because of its flexibility and sensitivity, this sensor can be used widely as a robotic skin in humanoid robots.

A biomimetic soft fingertip applicable to haptic feedback systems for texture identification

Humans recognize textures using the tactile data obtained from the human somatosensory system. Recognition of textures allows humans discriminate objects and materials. Moreover, by understanding the objects or materials texture, the human intuitively estimates roughness and the friction properties of the object or the material. This ability is necessary for object manipulative tasks. Likewise artificial haptic systems too, should have the ability to encode textures and feedback those data to haptic applications such as haptic displays. In this paper a biomimetic soft fingertip sensor that can be used in above haptic systems is introduced. The fingertip has the ability to detect force and vibration modalities. We propose three features calculated from the covariance signal of two adjacent accelerometers in the fingertip to use in texture identification. The covariance signal is transformed using Discrete Wavelet Transform (DWT) and the three features mentioned below are calculated. The mean and variance of the approximate signal, and the energies of the detailed signal are chosen as features. Then, the proposed features were validate by using those in an Artificial Neural Network (ANN) to classify seven wood samples. The results showed a 65% success rate in classifying wood samples and that the proposed features are acceptable to encode textures.

Analysis of sliding of a soft fingertip embedded with a novel micro force/moment sensor: Simulation, experiment, and application

We have investigated the deformation of a soft fingertip when it slides. This process was first simulated using the non-linear Finite Element Analysis (FEA) method. Based on the results of this simulation, we designed experiments to observe the sliding and object grasping of a soft fingertip, in which a 3-DOF (degree of freedom) micro force/moment sensor was embedded. With this sensor, forces and moments acting in the fingertip are measured based on the piezoresistive effect. These measurements provide information on the status of contact and sliding of a soft fingertip on a surface. Based on these results, incipient slip, which has an important role in object gripping by a robot manipulator, can be realized. Textiles texture recognition experiments were also conducted to assess potentials of the fingertip in tactile and texture perception.

Experimental investigation of surface identification ability of a low-profile fabric tactile sensor

Humans usually distinguish objects by sliding their fingertips on the surface to feel the texture via mechanoreceptor underneath the skin. We have developed a human-imitated system for robotic fingertip to sense objects texture via sliding action. Design of the sensory skin was inspired by the localized displacement phenomenon of a sliding soft fingertip ([1]) to capture stick-slip events on the contact surface that mainly represent texture characteristics. The soft skin is knitted by electro-conductive tension-sensitive yarns, then covered over a hemispherical fingertip. The pile-shaped surface of the fabric sensor enhances tangential traction detection ability of the sensor, even though the normal load is also sensible. Our aim is to exploit this sensor in applications regarding relative sliding between the touched object and the surface of the sensor, such as slip detection ([2]), and surface identification in this paper. In surface encoding, we have experimentally investigated ability of the fabric sensor in recognition touched objects via multiple machine learning algorithms, such as naive Bayes, Multi-Layer Artificial Neural Network (ANN) with input extracted from autoregressive models, and ANN with input extracted from Discrete Wavelet Transformation (DWT), have been trained to distinguish three typical textures. As a result, we have found that the last method outperforms the remains with an average successful rate of 90%.

Modeling and Analysis of a Frictional Sliding Soft Fingertip, and Experimental Validations

We have proposed a dynamic model of a soft fingertip to investigate its sliding motion on a plane with friction. The fingertip is comprised virtually of a finite number of elastic, compressible and bendable cantilevers whose free ends act as contact points. Moreover, we utilize virtual linkage spring–damper elements between the contact points to represent interactions between neighboring cantilevers on the contact surface. By introducing Coulombs law into each contact point, we are able to capture the frictional characteristic during sliding motions of the fingertip, especially stick–slip motions. We also present experimental results to validate this model.

Understanding Slip Perception of Soft Fingertips by Modeling and Simulating Stick-Slip Phenomenon

Slip, especially incipient slip, is a complicated process for soft fingertips; and detection of this slip is important for stable manipulations by both human and robotic fingertips. Experimental research has attempted to explainthis phenomenon, but dynamic changes during this process could not be fully delineated. We propose here a dynamic model to investigate the sliding motion of soft fingertips on a plane with friction. The fingertip is comprised of a finite number of elastic, compressible and bendable cantilevers whose fr ee ends act as infinitesimal contact points. The contact surfac e is afterward meshed using a finite element method based on the coordinates of the contact points. By introducing Coulomb’ s law and contact compliance into each contact point, we were able to assess the frictional characteristics of the sliding motio ns of the fingertips. We also could successfully describe the dynamic ally localized displacements on the contact surface during stic k-slip transition, displacements that represent the sliding moti on of a soft fingertip. This model can be applied to different shape s of robotic fingertip, including the cylindrical and hemispherical models tested here. We also performed experiments to valida te the proposed simulation, including force/moment and vision setups. I. I NTRODUCTION Recent research in robotics has focused on the dexterous manipulation of objects using soft fingered robotic hands, e specially anthropomorphic hands. This type of research can b e categorized into two main groups. The first consists of studi es that focus on analyzing the contact mechanics between soft fingers and objects [1]. In the second, tactile sensing syste ms imitating those of humans, along with many types of sensors, have been developed to simulate human abilities in object grasping and dexterous handling [2]. Whereas the former studies consisted primarily of analysis of stable grasping or object postures controlled by soft fingertips during a pushi ng or rolling motion on the surface of objects; the latter studi es concentrated on the tactile texture perceptions of sensory fingertips, to increase efficiency during object manipulati on processes. Among the various movements of fingertips during object handling is slide/slip, which often occurs during co ntact and is considered important in dexterous manipulation [3]. For example, to assess the texture of an object’s surface, the fingertip needs to slide slightly on the surface to extrac t information about its roughness or friction. The sliding mo tion of an object between fingertips during grasping, or incipien t slip, is a crucial factor in stable object manipulation. How ever, while the latter studies have addressed all types of fingerti p motions, including pressing, rolling, and rubbing; the for mer studies have focused primarily on pure pushing and rolling movements, while ignoring slip or slide. The difficulties in modeling the sliding motion of fingertips comes from their compliance, their partial movements on contact surfaces, a nd the force/moment of friction. Kao and Cutkosky [4] have proposed a method that combines compliance and friction on a limited surface to compute the relative sliding motion between a grasped object and soft fingers. That article showe d concrete results in modeling contact and in approximation o f gross-motion planning. Fearing [5] introduced an algorith m for automatic stable grasping of polygonal objects by two fingers and a point contact with friction. Research has also addressed pushing operations using manipulators. Yoshika wa et al. [6] proposed a pushing method to identify the center of friction of an object with an unknown distribution of fricti on. In a similar method [7], relevant friction parameters were estimated by performing experimental pushes and observing the resultant motion. Nevertheless, most previous researc h has dealt only with quasi-static analysis, including gross sli ding, not on stick-slip phases (i.e. how and when it occurs), or par tial slip on the contact surface of a soft fingertip. In contrast, some contact models have been used in simulation, such as analytical ([8]) and conventional penalty methods ([9]) . Although the former methods result in a fast and accurate description of contact, they limit penetration between obj ects; in contrast, the latter methods can identify the contact are a, but can only use penetrative contacts. At the same time, the Fini te Element Analysis (FEA) model of a contacting soft fingertip was also developed. This method, however, is time-consumin g and usually solved in a static field ([10]). In contrast to the above approaches, Barbagli et al [11] attempted to determine the model that best fit the real rotational friction properti es of human fingertips and then extended the object/proxy algorit hm to simulate soft finger contact. In this paper, we have attempted to determine the frictional characteristics of a 3-D soft fingertip during unilateral sl iding motion relative to an object. We have proposed a method to model vertical and horizontal deformations of the soft fingertip during sliding by introducing virtual cantilever s. Moreover, we have modeled the contact surface by employing the FEA method, along with Coulomb’s frictional law. This can significantly reduce calculation time, while still assu ring he dynamic behavior of the system during stick-slip transi tion. Non-contact cantilevers Contact cantilevers Contact surface Contact cantilevers (a) Modeled fingertip (b) Meshed contact surface Tp Pi k Pj P Fig. 1. Model of a sliding soft fingertip with virtual cantile v rs and meshed contact surface. Results from simulation and experimental validation can th eoretically explain, for the first time, localized movements on the contact surface during the stick-slip phase, as determi n d experimentally. II. PROPOSEDHYBRID MODELING OF A HEMISPHERICAL SOFT FINGERTIP Many shapes are used to form robotic soft fingertips. We have categorized them into two main groups with respect to grasped objects with rigid flat facets. The first generates uniform distribution of normal stress, such as square and rectangular fingertips. The second is characterized by the n onuniform distribution of contact force, including cylindri cal, hemispherical, and human-like fingertips. Both groups are distinguished by rigid fingertips in compliant contact, fea turing a pre-slide stage and characteristics of hysteresis. Mo re particularly with the second group, the pre-slide generate s localized displacements on the contact surface, which are important in incipient slip detection ([20]). A dynamic mod el with the appearance of friction is required to represent the se phenomena and characteristics of soft fingertips. We have developed a model, which, with some simplifications, can be employed to investigate slip motions of soft fingertips of different shapes, especially during stick-slip transitio n. Since this model focuses onwhat happenson the entirecontact surface, we have combined discrete and FEA methods to model a general 3-D soft fingertip. By so doing, the calculati on times will be reduced considerably, while still assuring th e correctness of the model. Assume that a three dimensional (3-D) soft fingertip with an arbitrary but continuous outer surface is pushed vertica lly with a specific contact penetration by a normal force Fn on a rigid flat plane, and slides horizontally with an external fo rce Ft (Fig. 1(a)). Inoue ([1]) proposed a soft fingertip model comprised of an infinite number of vertical elastic virtual springs that could be used to investigate the deformation of the fingertip during pushing or rolling motion on an object. This model, however, was not sufficient to demonstrate slidi ng motion with the appearance of frictional force. Therefore, instead of virtual springs, we have proposed a model, in which the soft fingertip is comprised of a finite number of virtual elastic cantilevers that are compressible, tensil , and bendable (Fig. 1(b)). This model can represent the diverse deformations of a soft fingertip during a sliding motion, in which the fingertip is pushed and slid at the same time. These cantilevers are fixed on the flat surface of the fingertip, with their free ends on its outer surface. Each cantilever has a uniform circular cross sectional area, whereas the lengths of the cantilevers differ depending on their coordinates with in the fingertip. When the fingertip is pushed, some cantilevers will make contact with the object, causing the deformations of these cantilevers. Based on geometrical distributions, we were able to calculate the deformation of each contacting cantilever, which have non-uniform distribution on the con tact surface. Afterwards, we meshed the 2-D contact surface usin g the Voight model to describe the elastic and viscous propert ies of the contact surface ([12]) on the O −XY coordinates. Each node was equivalent to the free end of a contacting cantileve r. Each element is a triangle Tp with three nodes referred as Pi, Pj , andPk in a counterclockwise direction. Fig. 1(c) illustrates a possible mesh of the contact surface. For the sake of simplicity, we made three assumptions: 1) When a cantilever is bent, its deformation is significant only at the free end. 2) Interactions between continuous cantilevers only occur between their free ends on the contact surface. 3) Only cantilevers whose free ends are acting on the contact surface are considered (dark colored cantilevers in Fig. 1(b)). Cantilevers outside the contact surface are deemed irrelevant to the sliding motion of the fingertip (light colored cantilevers in Fig. 1(b)). As a result, when the fingertip is pushed and slid, its deforma tion will be represented by deformations of all contacting c antilevers. Moreover, external forces acting on each node on t he contact surface can be assessed by calculating the compress ing and bending forces of the corresponding can