Robust nonlinear control of uncertain systems: An application to intelligent vehicle highway systems (IVHS)

One of the major challenges in the design of controllers for dynamical systems is the imperfect knowledge of the system. In this dissertation we present a theoretical investigation of robust control techniques for a class of uncertain systems. A systematic application of the theory is then made to the longitudinal control of a platoon of vehicles in an Intelligent Vehicle Highway System.; Assuming an a priori knowledge of the bounds on the systems uncertainties the control problem can be solved using deterministic methods. In particular, we utilized sliding control techniques. However, these techniques require strict structural conditions on the uncertainty known as the matching conditions. Consequently, we present a modified technique which utilizes synthetic inputs and a multiple sliding surface formulation. The methodology, developed for systems transformable into computable normal form as well as nonlinear forms, guarantees uniform ultimate boundedness of the tracking error.; As a prerequisite to developing a solution to the longitudinal control problem we developed a nonlinear longitudinal vehicle model. The model possessed a sufficient degree of fidelity to accurately predict the subtleties of an actual vehicle. Singular Perturbation and linearization techniques were then utilized to justify model simplifications for the controller based models. Regions of applicability for each of the models were validated through simulation and comparison with actual vehicle test data.; We then considered the problem motivated by the longitudinal control of a platoon of autonomous vehicles. In the attempt to study this problem we were faced with several problems, namely a noninteracting control problem, internal dynamics, and a systematic way of analyzing vehicle and platoon string stability. The first two issues were first formulated in a general context and then specific solutions were then discussed. The latter two were considered within the Composite System (CS) Method often used in large scale system analysis. Stability for each subsystem was defined in terms of its ability to track a particular desired trajectory. Stability of the interconnected systems or string stability was determined by the relative amplification (unstable)/attenuation (stable) of each subsystem's spacing error as the string index increases. The result provided was the determination of controller with sufficient complexity for acceptable nominal and robust closed loop vehicle and platoon performance.; The control strategies were implemented experimentally using the Integrated Platoon Control System (IPCS). Early results using two and four car platoons successfully demonstrated the feasibility of the platoon concept using existing technologies. Recent test results focused on the determination of control model structural complexity previously studied via simulation. Experimental results confirm the results obtained through simulation as well as the applicability of the developed theory.