DocumentCode
3336854
Title
Hand control by a neural network using tactile and positional information
Author
Sperduti, Alessandro ; Starita, Antonina
Author_Institution
Dipartimento di Inf., Pisa Univ., Italy
fYear
1991
fDate
19-22 June 1991
Firstpage
1747
Abstract
The development of manipulators with perceptive tactile capabilities which are really similar to those of the human hand has been hindered until now by the complexity of the sensory structures involved. Furthermore, sensors and actuators need to be closely integrated. The paper considers contact touch, a particular perceptive capability, during tasks that don´t require force; and studies how it can be used by the actuators to control the movements of the hand. The organization that the authors propose is implemented with a structured neural network. Backpropagation is used to map tactile receptors onto motor actuators.<>
Keywords
backpropagation; manipulators; neural nets; position control; tactile sensors; contact touch; hand control; manipulators; motor actuators; neural network; perceptive tactile capabilities; position control; positional information; tactile receptors; Actuators; Backpropagation; Computational modeling; Computer networks; Force control; Force sensors; Humans; Neural networks; Quantization; Tactile sensors;
fLanguage
English
Publisher
ieee
Conference_Titel
Advanced Robotics, 1991. 'Robots in Unstructured Environments', 91 ICAR., Fifth International Conference on
Conference_Location
Pisa, Italy
Print_ISBN
0-7803-0078-5
Type
conf
DOI
10.1109/ICAR.1991.240350
Filename
240350
Link To Document