Premium
Stability analysis of nonholonomic extremum seeking systems via Lie brackets approximations
Author(s) -
Grushkovskaya Victoria
Publication year - 2021
Publication title -
pamm
Language(s) - English
Resource type - Journals
ISSN - 1617-7061
DOI - 10.1002/pamm.202000024
Subject(s) - nonholonomic system , controllability , rank condition , control theory (sociology) , rank (graph theory) , mathematics , stability (learning theory) , component (thermodynamics) , function (biology) , lie group , convex hull , regular polygon , computer science , control (management) , robot , mobile robot , artificial intelligence , pure mathematics , physics , geometry , combinatorics , machine learning , evolutionary biology , biology , thermodynamics
Abstract The paper presents a novel family of dynamic extremum seeking controls for nonholonomic systems with output described by a convex function. It is assumed that the vector fields of the system satisfy the controllability rank condition with first‐ and second‐order Lie brackets. The proposed algorithm consists of two components. The first one is the stabilizing model‐based component constructed based on control strategies for nonholonomic systems. The second one is the extremum seeking model‐free component relying only on the values of the output of the system. The resulting control strategy optimizes the system behavior in the sense of minimizing its output. Namely, the main result of the paper states practical exponential stability of the optimal point, at which the output takes its minimal value.