Optimal output tracking control for nonlinear systems via successive approximation approach (Q869828)

The desired reference output of a control system is presented as a solution to a completely observable asymptotically stable linear system of ordinary differential equations. The authors propose a procedure of designing tracking control. The problem of tracking is thus formulated as an optimal control problem with a quadratic functional, but nonlinear dynamics. The necessary conditions for optimality lead to a two-point boundary-value problem for the state and adjoint systems of equations. For a linear problem, such a boundary-value problem can be resolved through a matrix Riccati equation. To employ this feature of a linear system in the solution of the problem under consideration, the right-hand side of the state system of equations is transformed into a sum of a linear term and a nonlinear remainder term using MacLaurin expansion. To solve the two-point boundary-value problem, the authors apply the method of sequential approximations. The main result is that the successive approximations (Picard's iterations) uniformly converge to the optimal solution. This result has been obtained for both finite and infinite horizon systems. The method is constructive and has been illustrated on an 2D example. The example, however does not demonstrate full benefits of the proposed procedure since it is based on a single-input tracking control system with no nonlinear term.