Output-feedback global tracking for unknown control direction plants with application to extremum-seeking control

53rd IEEE Conference on Decision and Control

Aminde

Hsu

2014

Self Cite

In this paper, we consider the problem of seeking a light source with a cart-track linear positioning servomechanism. The cart does not have the capability of sensing its position/velocity or the position of the source but is capable of sensing the light signal generated by a light bulb. The light field created by the light source decays away from the position of the light bulb but the servo does not have the knowledge of the functional form of this field. We employ an extremum seeking control strategy based on monitoring functions recently proposed by the authors. First, we briefly introduce the theoretical results of the monitoring function based extremum seeking controller for one-dimensional (1D) optimization. The class of systems previously considered was restricted to the relative degree one case. Here, as a theoretical contribution, generalization is achieved to include more general dynamics with arbitrary and uncertain relative degrees. In lieu of using differentiators to compensate the relative degree, a time-scaling procedure is developed. Then, we present the design for a light-seeking servomechanism and obtain experimental results of such a mounted device performing localizing and tracking of the level-sets of the light source.

Section: B Basic Error Equationmentioning

confidence: 98%

Section: A Control Objective and Main Assumptionsmentioning

confidence: 99%

Section: B Time-scaling For Control Redesignmentioning

confidence: 99%

“…The time-scaling (32), allow us to consider the original plant (20)- (22) in a different time-scale being controlled by the controller (14) properly scaled by µu, see (33).…”

Section: Scaled Controller Parametersmentioning

confidence: 99%

Section: Uncertain and Arbitrary Relative Degreesmentioning

confidence: 99%

See 3 more Smart Citations

Monitoring function based extremum seeking control for uncertain relative degrees with light source seeking experiments

53rd IEEE Conference on Decision and Control

Aminde

Hsu

2014

Self Cite

Binary robust adaptive control with monitoring functions for systems under unknown high‐frequency‐gain sign, parametric uncertainties and unmodeled dynamics

Adaptive Control & Signal

Peixoto

Nunes

2015

Self Cite

In this paper, we propose an adaptive output-feedback controller for uncertain linear systems without a priori knowledge of the plant high-frequency-gain sign. To deal with parametric uncertainties and unmodeled dynamics, we consider a robust adaptive strategy named binary model reference adaptive control, which has the good transient properties and robustness of sliding mode control with the important advantage of having a continuous control signal free of chattering. The effective way of tackling unknown high-frequencygain sign is employing monitoring functions. The developed adaptive control guarantees global exponential stability of the closed-loop error system with respect to a compact residual set. Moreover, in the absence of unmodeled dynamics, exact tracking of a reference signal can be achieved. Numerical simulations illustrate the efficacy of the proposed approach.

Multivariable extremum seeking control via cyclic search and monitoring function

Aminde

Adaptive Control & Signal

Hsu

2020

Self Cite

Summary This article addresses the design of a multivariable extremum‐seeking controller for a class of nonlinear static multi‐input systems. A method is proposed based on cyclic search directions and a monitoring function to guarantee the objective function to track an adequate reference trajectory via sliding‐mode control. The method allows achieving an arbitrarily small neighborhood of the desired optimal point (extremum) starting from any initial conditions. Simulation results illustrate the significant advantages of the proposed multivariable control strategy in terms of speed of convergence, global convergence and small residual errors.