Development of software tools was never primary aim of our research, however development of methodologies and algorithms is impossible without proper software support. At present, we are dealing with increasingly more complex systems, hence requirement on reliability and flexiblity of software tools are growing. The following projects are actively developped and maintained:
Mixtools is a toolbox designed for learning, prediction and control design with probability mixtures. It makes the firm basis of the other software products.
This project is a successor of Object-Oriented toolbox Mixtools 3000. Decision making is interpreted as a method of dedicated objects: decision makers. The library contains a range of common decision-makers, such as estimators, Bayesian filters (including particle filters, Kalman filters, etc.) Decision makers for control, (i.e. LQG controller) are under development.
GPC toolbox is a toolbox for MATLAB and Simulink. It serves for obtaining the basic knowledge about the Generalized Predictive Control (GPC) applied to linear single-input single-output systems, which are described by input-output differential equation or state-space form.
A set of Matlab routines developed for analysis of medical image sequences. Many methods are general purpose blind source separation tools and can be used with any data. Other are specific to medical image sequences.
Jobcontrol should help the user to focuse on the application problem at hand without spending too much energy on proper combination of tools from Mixtools into the problem solver.
Designer is a label for a complete chain of algorithms that serve for a complete computerized design of adaptive controllers. They start from raw data measured on a real system, user's knowledge and wishes and result into a completely pre-tuned aand verified controller.
LQ toolbox is a MATLAB - Simulink based toolbox for design of LQ controllers based on input-output models and simulation of examples with single input and single output systems. Fixed model and adaptive versions are supported.
Responsible for information: AS