Exploiting the virtue of redundancy

作者: A. Karniel , R. Meir , G.F. Inbar

DOI: 10.1109/IJCNN.1999.832731

关键词: Motor controlUncountable setMathematicsControl theoryArtificial intelligenceRedundancy (engineering)InverseLearning theoryTheoretical computer scienceCountable setObject (computer science)

摘要: Bernstein suggests that redundancy is the main reason for superb dexterity of human motor control. However, introducing in an inversely controlled object, results ill-posed problem. We suggest learning all possible solutions and choosing one them real time. In this paper we define differentiate between finite, countable uncountable redundancy. introduce a general concept multiple controller describe specific architecture, polyhedral mixture linear experts (PMLE). extend some notions theory to case valued functions stress difference estimated inverse estimation. Then show PMLE suitable serving as controller.

参考文章(4)
Gideon F. Inbar, Amir Karniel, Ron Meir, Polyhedral Mixture of Linear Experts for Many-To-One Mapping Inversion the european symposium on artificial neural networks. pp. 155- 160 ,(1998)
Mathukumalli Vidyasagar, A Theory of Learning and Generalization ,(1997)
A. Karniel, G.F. Inbar, Human motor control: learning to control a time-varying, nonlinear, many-to-one system systems man and cybernetics. ,vol. 30, pp. 1- 11 ,(2000) , 10.1109/5326.827449
Paul J. Werbos, W. Thomas Miller, Richard S. Sutton, Connectionist Learning for Control MIT Press. pp. 5- 58 ,(1995)