A pattern classification by dynamic tactile sense information processing

Kinoshita, G.; Aida, Shuhei; Mori, Masahiro

doi:10.1016/0031-3203(75)90009-6

Cited by 34 publications

(12 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Previous work in tactile sensing for recognition tasks has emphasized traditional pattern recognition paradigms on arrays of sensor data, similar to early machine vision work (Kinoshita, Aida and Mori 1975;Hillis 1982;Ozaki et al 1982;Overton 1984). Most sensing has been static in that the sensor is larger than the object and a single touch or &dquo;handprint&dquo; is used for recognition.…”

Section: Tactile Sensingmentioning

confidence: 97%

Integrating Vision and Touch for Object Recognition Tasks

Allen

1988

The International Journal of Robotics Research

151

View full text Add to dashboard Cite

A robotic system for object recognition is described that uses passive stereo vision and active exploratory tactile sensing. The complementary nature of these sensing modalities allows the system to discover the underlying 3-D structure of the objects to be recognized. This structure is embodied in rich, hierarchical, viewpoint-independent 3-D models of the objects which include curved surfaces, concavities and holes. The vision processing provides sparse 3-D data about regions of interest that are then actively explored by the tactile sensor mounted on the end of a six-degree-of-freedom manipulator.A robust, hierarchical procedure has been developed to integrate the visual and tactile data into accurate 3-D surface and feature primitives. This integration of vision and touch provides geometric measures of the surfaces and features that are used in a matching phase to find model objects that are consistent with the sensory data. Methods for verification of the hypothesis are presented, including the sensing of visually occluded areas with the tactile sensor. A number of experiments have been performed using real sensors and real, noisy data to demonstrate the utility of these methods and the ability of such a system to recognize objects that would be difficult for a system using vision alone.

show abstract

Section: Tactile Sensingmentioning

confidence: 97%

Integrating Vision and Touch for Object Recognition Tasks

Allen

1988

The International Journal of Robotics Research

151

View full text Add to dashboard Cite

show abstract

“…Tactile sensors have been constructed using piezoelectric, piezoresistive, magneto-electric, capacitive, optical, and other technologies (Ueda et al 1972;Kinoshita 1975;Stojiljkovic and Clot 1977;Purbrick 1980;Schneiter 1982;Larcombe 1983;Rebman and Trull 1983;Boie 1984;Raibert 1984;Seigel et al 1987;Tise 1988;Speeter 1988). The most common variables transduced by these devices are or, or c, , the normal components of stress and strain.…”

Section: Receptor Responsementioning

confidence: 98%

Three-dimensional Finite Element Analysis of Elastic Continua for Tactile Sensing

Speeter

1992

The International Journal of Robotics Research

View full text Add to dashboard Cite

This article presents a method for determining threedimensional stress and strain tensors within elastic layers covering or embedding tactile sensors. When an object comes in contact with the elastic surface of a sensing array, the distribution of forces on the surface is mechanically filtered before reaching the underlying sensing elements. In order to predict the response of embedded sensors and to understand the nature of the transduced variables, the relationship between surface forces and interior stress and strain tensors must be known. To determine this, we need to know the shape of the surface contact area and the distribution of forces over that area, both normal and tangential. The algorithm described here discretizes the known surface force distribution into a dense array of independent point loads, approximating the surface force profile. The stress field produced by a single normal and/or a single tangential point load can be expressed analytically, given the elasticity and compressibility of the material. Fields for each point load are calculated and can be summed to determine the resultant stress tensor at any point within the medium. Strains can be computed from stresses using a simple set of translation formulas.A detailed example is illustrated, analyzing the interaction between a sphere and a planar elastic half space. From the example the relative effects of depth and distance from the center of contact on the response of underlying sensors is discussed. Using results from the example, methods are proposed to exploit sensors at different depths and at different orientations within the medium. An algorithm for determining the direction of surface tangential forces and a method for determining the ratio between normal and tangential forces are analyzed, using unidirectionally sensitive sensors at different depths in the elastic layer.

show abstract

“…The object surface was explored, and objects were classified by local curvature measurements. Kinoshita [1977;Kinoshita, Aida and Mohri 1975] demonstrated discrimination of circular, square and triangular cylin-ders based on the total number of sensing sites activated, using 384 binary sensors on a 5 finger hand. Hillis [1981] proposed to classify six small parts based on 3 ''features" j the shape (object long or round), bumps, and whether or not the part rolled with finger motion.…”

Section: Shape Interpretation Without Specific Modelsmentioning

confidence: 99%

Dextrous Robot Hands

Venkataraman

Iberall

1990

View full text Add to dashboard Cite

PrefaceDextrous manipulation has been a topic of interest in industrial assembly, prosthetic hand design, and in the study of human movement. In recent years, dextrous robot hands, developed in the United States, Europe, and Japan, have become available as research tools. While significant progress is being made in their design, construction, and low level control, the true potential of dextrous mechanical hands has yet to be realized. Roboticists are up against fundamental problems in developing algorithms for solving real-time computations in multi-sensory processing and motor control.One possible way to develop better dextrous robot hands is through the study of human hands. After decades of research into the control problem of human arm movement, motor behaviorists are discovering key features of sensorimotor integration in the central nervous system. In a sense, they are reverse engineering the controller of a highly versatile, multiple degree of freedom, sensor-based 'machine', of the type that roboticists would love to build.The aim of this book is to explore parallels in sensorimotor integration in dextrous robot and human hands. While it is a view of the state of knowledge in 1989, we feel that one way the robotics community can design and learn to control more sophisticated robot hands is for them to work with motor behaviorists. As seen from some of the work presented in this book, such alliances have already proven fruitful.The significance of this text is that it brings together researchers in dextrous manipulation, and is a generally accessible resource for both experimentalists and engineers. In a computer science department, it would be a good companion text for a graduate level course in robotics, as it deals with integrating sensory information with motor control for performing a task. It would be a useful resource for a mechanical engineering course in robot hand design, as students could use it to gain insights into functional requirements for a dextrous hand. However, a strong feature of the book is that it could be used as a text for an interdisciplinary graduate course on sensorimotor integration for both experimentalists and engineers.This book grew out of the Workshop on Dextrous Robot Hands that occurred at the 1988 IEEE Conference on Robotics and Automation in Philadelphia, Pennsylvania. Highly successful, the workshop was attended by over 150 people, and was co-sponsored by the IEEE Computer Society and the Office of Naval Research.The first part of the book focuses on the functionality of human hands, discussing results from studies in prehension and apprehension. The first four chapters describe models of human hand function and present evidence VI Preface for why these models are valid. It is interesting to note that knowledgebased expert systems are described in Chapters 1 and 4 for codifying those models. The second part of the book focuses on control and computational architectures needed for dextrous hands. Chapters 5 through 7 describe architectural design decisions for three d...

show abstract

A pattern classification by dynamic tactile sense information processing

Cited by 34 publications

References 0 publications

Integrating Vision and Touch for Object Recognition Tasks

Integrating Vision and Touch for Object Recognition Tasks

Three-dimensional Finite Element Analysis of Elastic Continua for Tactile Sensing

Dextrous Robot Hands

Contact Info

Product

Resources

About