| A new control design approach for nonlinear systems is proposed. The design is based on a high-gain scaling approach utilizing a dynamic scaling parameter. The proposed approach features a number of advantages over previously available techniques. Firstly, the proposed methodology is very general and is applicable to both state-feedback and output-feedback problems for various classes of nonlinear systems including lower triangular (strict-feedback), upper triangular (feedforward), and certain non-triangular systems, thus providing a unified control design framework. Secondly, the design technique provides considerable robustness to parametric uncertainties, appended dynamics, and additive noise disturbances. The robustness to additive disturbances enables the first output-feedback control design for feedforward systems. The strong robustness properties also allow relaxation of two fundamental restrictions in previously available techniques for output-feedback control, the first restriction being that bounds on uncertain system terms cannot include cross-products between unknown parameters and unmeasured states and the second restriction being that appended dynamics should (in an ISS sense) be driven only by the system output. Both these restrictions are removed using the proposed approach. The generality of the proposed design philosophy also enables extensions (such as disturbance attenuation and handling of ISS appended dynamics) for feedforward systems along lines which have been hitherto investigated only for strict-feedback systems.; In the second part of the dissertation, two areas of controls applications are addressed: unmanned vehicles and large-scale supply chains. A detailed controls-oriented mathematical modeling of a sea surface vehicle is proposed along with some preliminary control designs. The modeling and control of a small helicopter are also addressed. A computationally tractable inventory control technique for large-scale bidirectional (reverse) supply chains is proposed along with an application to aircraft supply chains. It is expected that these problems could be promising application domains for the new control design approach developed in this dissertation. |
