The project studies Linear Complementarity System (LCS) modeling for piecewise affine systems with the goal of identification and observer design. This class of systems allows the description of complex model behavior, model non-differentiable dynamics arising in mechanical and electrical systems, and general function approximation without explicit state-space partitioning. Such an implicit representation is more compact than existing schemes, leading to a reduced set of parameters to be identified. An output-feedback control for these systems can also be described by an LCS with its dynamics dropping the need to explicit observer for each set, the design of these observers requires efficient solution to linear complementarity problem. To ensure the robustness of both the identified model and the observer, stability and set invariance should be imposed as requirements for the system’s identification and closed-loop operation.
Giorgio Valmorbida
Sorin Olaru
