Advanced Control of Robot Compliance Tasks Using Hybrid Intelligent Paradigms
In this paper, a new comprehensive intelligent control strategy based on connectionist learning is presented, effectively combining genetic algorithms(GA) with neural and wavelet classification. The proposed neural network classifies characteristics of environment, determines the control parameters for compliance control, and in coordination with basic learning compliance control algorithm, reduces the influence of robot dynamic model uncertainties. The effectiveness of the approach is shown by using a simple and efficient GA optimization procedures to tune and optimize the performance of a neural classifier and controller. Some compliant motion simulation experiments with robotic arm placed in contact with dynamic environment have been performed.
KeywordsMultilayer Perceptron Wavelet Neural Network Wavelet Network Neural Classifier Connectionist Learning
Unable to display preview. Download preview PDF.
- Katié, D., and Vukobratovic, M. (1998). A neural network-based classification of environment dynamic models for compliant control of manipulation robots. IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics 28 (1): 58–69.Google Scholar
- Vukobratovic, M., and Katié, D. (1994). Robot control structures for high-quality learning in flexible manufacturing tasks. In V.L.Plantamura, B.Soucek, and G.Visaggio., eds., Frontier Decision Support Concepts: Help Desk, Learning, Fuzzy Diagnoses, Quality Evaluation, Prediction, Evolution. New York: John Wiley & Sons,Inc,. 319–355.Google Scholar