Showing papers by "Petar V. Kokotovic published in 1993"
TL;DR: An adaptive backlash inverse controller for unknown plants with backlash is designed, a parametrization of the error caused by its estimate is given and the global boundedness of the closed-loop signals is proved.
181 citations
TL;DR: A new Lyapunov function is presented and used to design ‘softer’ control laws which exhibit the high-gain properties to a much lesser extent and achieve the same or better performance with less control effort.
155 citations
TL;DR: In this paper, the performance bounds for a class of adaptive and non-adaptive systems were derived and compared, showing that adaptation improves the overall performance without the undesirable effects of high gain.
137 citations
15 Dec 1993
TL;DR: In this paper, the authors consider nonlinear control systems for which an estimate x/spl circ/ of the system state x is available for feedback, and present conditions under which they can design a smooth feedback law u=/spl mu/(x /spl circ/) which renders the mapping from d/sub m/ to x globally input/output stable.
Abstract: We consider nonlinear control systems for which an estimate x/spl circ/ of the system state x is available for feedback. We assume x/spl circ/=x+d/sub m/, where d/sub m/(t) is an unknown locally bounded state measurement disturbance. We present conditions under which we can design a smooth feedback law u=/spl mu/(x/spl circ/) which renders the mapping from d/sub m/ to x globally input/output stable. For any initial condition, such a feedback law will guarantee that no finite escape times occur, that bounded disturbances d/sub m/ produce bounded signals, and that x/spl rarr/0 when d/sub m//spl rarr/0. We show that the class of systems for which such feedback laws exist include systems in strict feedback form. One important application is in the output feedback stabilization problem, where the disturbance d/sub m/ comes from a separately designed observer. >
76 citations
15 Dec 1993
TL;DR: Adaptive control schemes for systems with unknown backlash at the plant output are developed in this article, where a backlash inverse controller guarantees exact output tracking, and adaptive laws are designed to update the controller parameters and to guarantee bounded input-output stability.
Abstract: Adaptive control schemes for systems with unknown backlash at the plant output are developed. In the case of known backlash, a backlash inverse controller guarantees exact output tracking. When the backlash characteristics is unknown, adaptive laws are designed to update the controller parameters and to guarantee bounded input-output stability. Simulations show significant improvements of the system performance achieved by such adaptive backlash inverse controllers. >
58 citations
TL;DR: An adaptive algorithm for adjusting the gains of a vehicle speed control system is presented, which helps the design of a single speed control module that does not need additional calibration or sacrifices in performance for certain car lines.
Abstract: An adaptive algorithm for adjusting the gains of a vehicle speed control system is presented. By continuously adjusting the proportional-integral control gains, speed control performance can be optimized for each vehicle and operating condition. This helps the design of a single speed control module that does not need additional calibration or sacrifices in performance for certain car lines. It also allows improved performance for changing road conditions not possible with a fixed-gain control or other types of adaptive control. The results of initial vehicle testing confirm the performance improvements and robustness of the adaptive controller. >
45 citations
02 Jun 1993
TL;DR: In this article, an adaptive version of the backlash inverse is proposed for feedback control of a known linear plant with an unknown backlash at its input, and applied to feedback control with unknown backlash.
Abstract: This paper addresses one of the nondifferentiable nonlinearities which appear very often in industrial control applications: backlash. We first present a right inverse of backlash and then give one of its possible numerical implementations. Characteristics like backlash are seldom known, so an adaptive version of our backlash inverse is more suitable for applications. We develop an adaptive backlash inverse scheme and apply it to feedback control of a known linear plant with an unknown backlash at its input. We use simulation results to illustrate the achieved performance improvements.
43 citations
TL;DR: In this article, a pseudogradient adaptive algorithm is applied to self-tune a proportional integral (PI) puddlewidth controller for consumable-electrode gas metal arc welding.
Abstract: A pseudogradient adaptive algorithm is successfully applied to self-tune a proportional-integral (PI) puddle-width controller for consumable-electrode gas metal arc welding. The gradient of the output with respect to the controller parameters is approximated and used to form a steepest-descent algorithm to minimize the squared output error. Experimental data confirming the algorithm performance are presented. >
26 citations
TL;DR: In this article, a recursive design of adaptive controllers for linear systems is presented from a passivity perspective, where the adaptation loop for a minimum phase plant with known but arbitrary relative degree possesses a strict passivity property.
21 citations
TL;DR: In this article, a controller structure with an adaptive dead-zone inverse is proposed for plants with unknown dead-zones at output, which results in a linear parametrization which is crucial for developing adaptive update laws.
19 citations
15 Dec 1993
TL;DR: In this paper, the authors present a new approach to adaptive nonlinear control based on a complete controller-identifier separation which has long been a goal in adaptive system design, in which it is possible not only to separately design and analyze the parameter identifier, but also to independently guarantee properties achieved by the controller.
Abstract: We present a new approach to adaptive nonlinear control based on a complete controller-identifier separation which has long been a goal in adaptive system design. In this approach it is possible not only to separately design and analyze the parameter identifier, but also to independently guarantee properties achieved by the controller. Our controllers guarantee input-to-state stability with respect to the parameter error and its derivative as inputs. The parameter identifiers, in turn, guarantee boundedness of these signals. A key ingredient in the identifier design, as well as the convergence analysis, is a nonlinear generalization of the famous swapping lemma. The identifiers are based both on the plant model and the error system and allow a wide variety of update laws-gradient and least-squares, normalized and unnormalized. >
02 Jun 1993
TL;DR: A new design procedure for the nonlinear system transformable into the output-feedback canonical form is presented, using an observer-based parameter identifier which substitutes the tuning function as a tool for avoiding overparametrization.
Abstract: We present a new design procedure for the nonlinear system transformable into the output-feedback canonical form. The key ingredient is an observer-based parameter identifier which substitutes the tuning function as a tool for avoiding overparametrization.
15 Dec 1993
TL;DR: A new design procedure is developed for adaptive output-feedback control of nonlinear systems that guarantees boundedness even without adaptation and without knowledge of bounds on the unknown parameters.
Abstract: A new design procedure is developed for adaptive output-feedback control of nonlinear systems The procedure guarantees boundedness even without adaptation and without knowledge of bounds on the unknown parameters It also eliminates overparametrization using tuning functions Among other new features are the construction of a single Lyapunov function which incorporates all error variables, and the possibility for systematic improvement of transient performance >
15 Dec 1993
TL;DR: A state feedback control is presented which achieves asymptotic tracking of signals generated by a known exosystem for single-input single-output nonlinear systems with unknown parameters entering linearly.
Abstract: A state feedback control is presented which achieves asymptotic tracking of signals generated by a known exosystem for single-input single-output nonlinear systems with unknown parameters entering linearly The following assumptions are required: the nominal system is feedback linearizable; the tracking dynamics are not affected by unknown parameters; the linear approximation of the zero dynamics has no eigenvalue with zero real part; pure feedback conditions are satisfied on the vector fields multiplying unknown parameters This extends to a class of nonlinear uncertain systems the regulator theory of Isidori and Byrnes (1990) and provides adaptive tracking for systems with unbounded tracking dynamics which were excluded in Kanallakopoulos, Kokotovic and Morse (1991) >
01 Jan 1993
TL;DR: A backstepping procedure for designing non-dynamic feedback compensators for a class of uncertain nonlinear systems whose uncertainties vanish at an equilibrium point and one for systems with more general types of uncertainties.
Abstract: We present a backstepping procedure for designing non-dynamic feedback compensators for a class of uncertain nonlinear systems. We assume knowledge of nonlinear growth bounds on the system uncertainties, and we require that these bounds satisfy a strict feedback condition (a condition much less restrictive than the matching condition). We develop two design procedures, one for systems whose uncertainties vanish at an equilibrium point (in which case we achieve global uniform asymptotic stability) and one for systems with more general types of uncertainties (in which case we achieve global uniform ultimate boundedness).
25 May 1993
TL;DR: A comparison between the adaptive and the nonadaptive performance bounds demonstrates that adaptation improves the overall performance without the undesirable effects of high gain.
Abstract: Computable /spl Lscr/sub 2// and /spl Lscr/sub /spl infin// performance bounds are derived for a recently proposed class of adaptive systems which show that, in addition to global stability and asymptotic tracking, a systematic improvement of transient performance can be achieved. The underlying linear nonadaptive controller is shown to possess a parametric robustness property, but for large parameter uncertainty it requires high gain. A comparison between the adaptive and the nonadaptive performance bounds demonstrates that adaptation improves the overall performance without the undesirable effects of high gain.
01 Jan 1993
TL;DR: This approach abandon the traditional linear certainty-equivalence concept and treat the control of linear plants with unknown parameters as a nonlinear problem, and develops a recursive design procedure which guarantees strict passivity of the adaptation loop.
Abstract: This talk is a tutorial presentation of a new approach to adaptive control of linear systems. In this approach we abandon the traditional linear certainty-equivalence concept and treat the control of linear plants with unknown parameters as a nonlinear problem. We develop a recursive design procedure which guarantees strict passivity of the adaptation loop. The states of the resulting adaptive system converge to a manifold whose dimension is smaller than with any previous scheme. A simulation comparison with a standard indirect linear scheme shows that the new nonlinear controller achieves superior transient performance without an increase in control effort.
25 May 1993
TL;DR: In this paper, an adaptive friction compensation control scheme for linear systems with unknown frictions was developed for the positioning of a space-station solar array, where a friction model including kinetic, viscous, static and Stribeck friction components was parameterized for adaptive estimation.
Abstract: Adaptive friction compensation control schemes are developed for linear systems with unknown frictions. A friction model including kinetic, viscous, static and Stribeck friction components is parameterized for adaptive estimation and employed for two control schemes: one for a known, and the other for an unknown linear part. Adaptive friction compensation in positioning of a space-station solar array is investigated. Simulations show that the proposed adaptive control schemes promise to bring considerable improvements to system performance.