TY - JOUR
T1 - Adaptive synchronization of unknown heterogeneous agents
T2 - An adaptive virtual model reference approach
AU - Baldi, Simone
AU - Frasca, Paolo
PY - 2019
Y1 - 2019
N2 - This work deals with state synchronization of heterogeneous linear agents with unknown dynamics. The problem is solved by formulating the synchronization problem as a special model reference adaptive control where each agent tries to converge to the model defined by its neighbors. For those agents that do not know the reference signal that drives the flock, a fictitious reference is estimated in place of the actual one: the estimation of such reference is distributed and requires measurements from neighbors. By using a matching condition assumption, which is imposed so that the agents can converge to the same behavior, the fictitious reference estimation leads to adaptive laws for the feedback and the coupling gains arising from distributed matching conditions. In addition, the coupling connection is not scalar as in most literature, but possibly vector-valued. The proposed approach is applicable to heterogeneous agents with arbitrarily large matched uncertainties. A Lyapunov-based approach is derived to show analytically asymptotic convergence of the synchronization error: robustification in the presence of bounded errors or unknown (constant) leader input is also discussed. Finally, a motivational example is presented in the context of Cooperative Adaptive Cruise Control and numerical examples are provided to demonstrate the effectiveness of the proposed method.
AB - This work deals with state synchronization of heterogeneous linear agents with unknown dynamics. The problem is solved by formulating the synchronization problem as a special model reference adaptive control where each agent tries to converge to the model defined by its neighbors. For those agents that do not know the reference signal that drives the flock, a fictitious reference is estimated in place of the actual one: the estimation of such reference is distributed and requires measurements from neighbors. By using a matching condition assumption, which is imposed so that the agents can converge to the same behavior, the fictitious reference estimation leads to adaptive laws for the feedback and the coupling gains arising from distributed matching conditions. In addition, the coupling connection is not scalar as in most literature, but possibly vector-valued. The proposed approach is applicable to heterogeneous agents with arbitrarily large matched uncertainties. A Lyapunov-based approach is derived to show analytically asymptotic convergence of the synchronization error: robustification in the presence of bounded errors or unknown (constant) leader input is also discussed. Finally, a motivational example is presented in the context of Cooperative Adaptive Cruise Control and numerical examples are provided to demonstrate the effectiveness of the proposed method.
UR - http://resolver.tudelft.nl/uuid:a403eafe-8a63-4091-8826-642c1182d690
UR - http://www.scopus.com/inward/record.url?scp=85041924532&partnerID=8YFLogxK
U2 - 10.1016/j.jfranklin.2018.01.022
DO - 10.1016/j.jfranklin.2018.01.022
M3 - Article
AN - SCOPUS:85041924532
SN - 0016-0032
VL - 356
SP - 935
EP - 955
JO - Journal of the Franklin Institute - Engineering and Applied Mathematics
JF - Journal of the Franklin Institute - Engineering and Applied Mathematics
IS - 2
ER -