Solving challenging control problems in the aerospace and robotics fields requires efficient and numerically reliable computational methods implemented in high-quality robust numerical control software. Collaborative efforts to develop numerical algorithms and implement high-quality robust numerical software have been pursued in the framework of several international and national cooperation projects such as the EU Thematic Network NICONET (Numerics In COntrol NETwork), and several research grants from the German and Swedish Research Foundation. The software implemented within NICONET belongs to the control library called SLICOT, which includes over 400 user-callable computational routines implemented in Fortran 77. SLICOT is based on the standard numerical libraries BLAS and LAPACK.
Based on SLICOT we developed several powerful and user-friendly toolsets for MATLAB. which can be directly called via suitable MEX-function interfaces. The collection of MEX-functions has been completed with user-friendly M-function interfaces to form SR-Tools, a comprehensive set of tools for basic linear systems analysis and synthesis computations. SR-Tools represents a powerful addition to the standard control tools available in MATLAB. In long term efforts, unique toolboxes for Descriptor Systems, Periodic and Multirate Systems, Fault Detection, and LFT-based Uncertain Systems Modelling, have been developed as well. They go far beyond the state-of-the-art available in standard control software tools.
At the basis of all these software implementations lie highly efficient numerically reliable algorithms developed by our team. Our algorithmic developments focussed on the following main areas:
- model and controller reduction
- manipulation of polynomial and rational matrices
- descriptor systems analysis and synthesis
- design of fault detection and isolation filters
- periodic systems