Nonlinear and Structured Interpolation for Robust Control
Abstract
We have explored the problem of extending the linear H(infinity) theory to nonlinear systems, robust distributed parameter control, and the robust analysis and synthesis of controllers for systems in the presence of various classes of structured perturbations. We have employed operator theory, partial differential equations, optimization theory, and invariant theory to study the problem of utilizing visual information in a feedback loop and the closely related application of visual tracking. We have considered a combination of methods from robust control and computer vision for this purpose. Our approach to computer vision and image processing is based on certain novel curvature dependent evolution equations that may be employed for image enhancement, denoising, active contours, edge detection, morphology, shape recognition, shape-from-shading, stereo disparity, and optical flow. We have combined this with our longstanding work in robust control, and applying these ideas to certain benchmark problems in visual tracking.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1998
- Accession Number
- ADA358498
Entities
People
- Allen R. Tannenbaum
Organizations
- University of Minnesota