Implementation of Hardwired Distributive Tactile Sensing for Innovative Flexible Digit

This paper focuses on the design of an efficient automated data interpretation system that is able to output its interpretation directly to describe the nature of contact with the tactile surface. The work integrates distributive tactile sensing technology, artificial neural networks (ANNs), and advanced FPGA-based application-specific real-time digital signal processing, to produce working systems for use in a steerable endoscope that retrieves shape and 'touch' information. Different configurations of the signal interpretation such as single and cascade neural networks were assessed. It was shown that the cascaded ANN architecture could achieve an overall accuracy of better than 94% when measuring clinically significant contact information. This technology has the potential to discriminate contact and palpation in minimal invasive surgery (MIS) tools.

[1]  Shawki Areibi,et al.  A Handel-C implementation of the back-propagation algorithm on field programmable gate arrays , 2005, 2005 International Conference on Reconfigurable Computing and FPGAs (ReConFig'05).

[2]  B.K.Y. Tam.  A novel actuated digit with tactile feedback for clinical applications , 2006 .

[3]  Francisco José Ballester-Merelo,et al.  Artificial neural network implementation on a single FPGA of a pipelined on-line backpropagation , 2000, Proceedings 13th International Symposium on System Synthesis.

[4]  Heekuck Oh,et al.  Neural Networks for Pattern Recognition , 1993, Adv. Comput..

[5]  P N Brett,et al.  A flexible digit with tactile feedback for invasive clinical applications , 2004, Proceedings of the Institution of Mechanical Engineers. Part H, Journal of engineering in medicine.

[6]  Mahdi Tavakoli,et al.  Design issues in a haptics-based master-slave system for minimally invasive surgery , 2004, IEEE International Conference on Robotics and Automation, 2004. Proceedings. ICRA '04. 2004.

[7]  R. R. Leal Ascencio,et al.  The study of a prototype of an artificial neural network on a field programmable gate array as a function approximator , 1999 .