Please use this identifier to cite or link to this item:
Type: Artigo de periódico
Title: Simple Auto-adaptive Neural Circuit For Control Of Human Gait: A Simulation Based On Back-propagation
Author: Sepulveda F.
Cliquet Jr. A.
Abstract: The acquisition of connectivity patterns in an artificial, three-layer neural circuit for control of human gait was simulated through a backpropagation algorithm. Input signals for the neural network were equivalent to hip, knee, and ankle angles, and to vertical ground reaction forces during walking. Neural network outputs consisted of signals proportional to the activity of five lower limb muscles. Many network input configurations and connectivity restrictions were tested. Data from normal gait were used for network training. Preliminary testing of network responses in unknown environments was effected by presenting the neural networks with actual pathologic data at the input layer and by later investigating network outputs. Results favored restricted connectivity between middle and output layer neurons. Further, explicit recall of recent events was found to improve network behavior.
Editor: ASME, New York, NY, United States
Rights: fechado
Identifier DOI: 
Date Issue: 1994
Appears in Collections:Unicamp - Artigos e Outros Documentos

Files in This Item:
There are no files associated with this item.

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.