Adaptive and Self-organizing Control Laboratory

Adaptive and Self-organizing Control Laboratory focuses on the development of a new topology preserving self-organizing neural networks. Our new neural networks are able to control the feature map probability according to a given target feature map probability for unknown and non-uniform input distributions. The integrally distributed adaptive learning laws were developed and analyzed for one-dimensional input space. The convergence properties of the new learning algorithms are analyzed using the Ljung's ODE approach and verified by simulations. Extensions to the multi-dimensional input space are currently under investigation.

This new self-organizing algorithms can be applied to several interesting research problems including:

•  A coordination of groups of distributed mobile sensing agents to control the usage probability for unknown and non-uniform distribution of location of signals.

•  A multi-resolution adaptive quantizer.

•  Local homeomorphism learning from an open subset of a manifold to the local coordinates.

Students:
Jongeun Choi

Publications:

Jongeun Choi and Roberto Horowitz, “A New Self-organizing Coordination Algorithm for Tunable Sensing Networks”, Submitted to the Panel of Intelligent Systems under Dynamic Systems and Controls, 2005 ASME International Mechanical Engineering Congress and Exposition November 6-11, 2005 Orlando, Florida USA.

Jongeun Choi and Roberto Horowitz, “A New One-Dimensional Coordination Algorithm for Distributed Sensing Networks”, Submitted to 44th IEEE Conference on Decision and Control and European Control Conference ECC 2005, Seville , Spain December 12-15 2005 .

Jongeun Choi and Roberto Horowitz, “Topology Preserving Neural Networks that Achieve a Prescribed Feature Map Probability Density Distribution” To appear in the 24th American Control Conference (ACC) June 8 to 10, 2005 , Portland , Oregon , USA .

R. Horowitz and L. Alvarez, "Convergence Properties of Self-Organizing Neural Networks," Proceedings of the IEEE International Conference on Neural Networks, Vol. 1 pp. 7-12, June 1996.

R. Horowitz and L. Alvarez, "Convergence Properties of Self-Organizing Neural Networks," Proceedings of the 1995 American Control Conference , pp. 1339-1344, Seattle , WA , June 1995.