Quadratic forms and approximate feed back linearization in discrete time

From MaRDI portal





The authors consider nonlinear, single-input, discrete-time control systems of the form \[ x(k+1)= F\bigl(x(k), u(k)\bigr) \] subject to a static or a dynamic feedback. Two such systems are referred to as statically quadratically feedback equivalent, if they coincide up to (and including) quadratic terms in the Taylor expansion of the map \(F\) after applying a state diffeomorphism and a feedback of the form \[ x= \xi+ \varphi (\xi), \quad v=u+ \alpha (\xi)+ \beta (\xi)u +\gamma u^2, \] where \(\alpha,\varphi\) are homogeneous functions of degree 2, \(\beta\) is homogeneous of degree 1, and \(\gamma\) is a constant. Characteristic numbers are introduced being complete invariants of the quadratic equivalence. As main result, two equivalent normal forms of discrete-time control systems under quadratic feedback equivalence have been discovered. A dynamic quadratic feedback equivalence has been defined and an established result states that discrete-time systems having controllable linear approximation are dynamically quadratically equivalent.











This page was built for publication: Quadratic forms and approximate feed back linearization in discrete time

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4348231)