Using switching linear systems to reach a prespecified target

Abstract EN

A conventional enumeration approach, and a new approach for the solution of the control of switched linear systems with input constraints are presented.

The main objective in this work is to determine control sequences {U‘(K), i -l , M and k –O, l, ….., N - l} which transfer the system from a given initial state X(0) to a specific target state Xt, (or to be as close as possible).

Considering both the conventional enumeration and the new approach, extensive computer simulations are performed using the discrete time system obtained by sampling (or discretizing) a continuous system.

The new approach is found to be more efficient than the enumeration one in terms of computations and computer storage and performs adequately under a variety of input data The procedure developed can be generalized and used to solve several versions of the switching control problem.

In particular, a procedure using a quadratic cost function (distance) is given for problems with time-invariant coefficients.