W.S. McMath

High sampling resolution tactile sensor for object recognition

The authors discuss a tactile sensor system with high sampling resolution and its application for the active perception of 3-D objects which are larger than the tactile probe dimensions. The experimental tactile sensor is based on force sensitive transducer technology. An elastic overlay with protruding tabs provides a de facto spatial sampling which reduces tactile sensor distortions caused by deformations of the overlay during probing. Successive images from the tactile probe are integrated with the kinesthetic parameters of a robot arm, producing an extended image from which object edges are extracted and used for model-based object recognition.<<ETX>>

Two-dimensional position recovery for a free-ranging automated guided vehicle

An absolute position recovery method for an automated guided vehicle (AGV) which can move in any direction (free-ranging) on the floor using vision is introduced. The floor is permanently encoded with the terms of a pseudorandom binary array requiring only one bit of code per quantization interval, independent of the desired resolution. >

Tactile pattern recognition using neural networks

This paper discusses an experimental tactile sensor based on force sensitive transducer technology and its application for pattern recognition. A multilayer neural network using the backpropagation algorithm was trained to recognize tactile images of letters embossed on wooden blocks.<<ETX>>

Three dimensional object recognition using integrated robotic vision and tactile sensing

Summarizes current development of a robotic sensing system to enhance the perceptive intelligence of manipulators working in relatively structured environments. The system consists of vision and tactile sensors which would be used to identify objects within a workspace and to determine their absolute position and orientation. A visual object recognition strategy is discussed, and a technique for fusing robot kinematics and tactile image data is described.<<ETX>>

Tactile sensor system for object recognition

This paper discusses equipment development and tactile image recognition aspects of an experimental robotic tactile sensing system.<<ETX>>

Tactile sensor for geometric profile perception

An experimental robotic tactile sensing system is under development to characterize and identify three-dimensional objects. The authors describe the experimental setup, which consists of a tactile sensor probe and a passive-compliant wrist mounted at the end of a manipulator arm. An incremental strategy for active sequential exploration of an object surface is discussed. A data function technique is presented, which combines information from the tactile sensor, the wrist and the kinematics of the robot, for precise object identification. Preliminary experimental results are discussed.<<ETX>>

Tactile sensing for space robotics

The authors discuss the state of the art in tactile sensing technology and the ability of current sensors to meet the harsh environmental demands of space applications (e.g. the Space Station). The performance of the tactile sensor, both as a stand-alone feedback instrument and in cooperation with visual feedback, is discussed. The authors describe the implementation and perception problems associated with the development of an experimental matrix-type tactile sensor using force-sensitive transducers.<<ETX>>

A robotic tactile perception system concept

An experimental robotic tactile system is under development to characterize and identify three-dimensional (3-D) objects through a procedure of active sequential exploration. The experimental setup is described, and a data fusion technique for precise object identification is presented. Preliminary test results demonstrate that overlapping tactile exploration allows recovery from sensor/target misalignment, enables improved data correlation, and consequently produces a more precise 3-D object identification. Only quasi-static (static kinesthetic plus robot kinematic) tactile sensing has been considered, with the objective of better establishing global 3-D geometric features of the target object.<<ETX>>

Model-based tactile object recognition

The authors discuss encoding and system development aspects for the model-based tactile recognition of 3D objects having their surfaces Braille-like marked with pseudo-random codes.<<ETX>>