TY - JOUR
T1 - Riccati-based design of event-triggered controllers for linear systems with delays
AU - Borgers, D.P.
AU - Dolk, V.S.
AU - Heemels, W.P.M.H.
PY - 2018/1/1
Y1 - 2018/1/1
N2 - In event-triggered control (ETC) systems, the measured state or output of the plant is sent to the controller at so-called event times. In many ETC systems, these event times are generated based on a static function of the current state or output measurement of the system and its sampled-and-held version that is available to the controller. Hence, the event-generator does not include any dynamics of its own. In contrast, dynamic event-generators trigger events based on additional dynamic variables, whose dynamics depend on the state or output of the system. In this paper, we propose new static and dynamic continuous event-generators (which require continuous measuring of the plant output) and periodic eventgenerators (which only require periodic sampling of the plant output) for linear control systems with communication delays. All event-generators we propose lead to closed-loop systems which are globally exponentially stable with a guaranteed decay rate, L2 -stable with a guaranteed L2 - gain, and have a guaranteed positive minimum inter-event time. By using new Riccati-based analysis tools tailored to linear systems, the conservatism in our decay rate and L2 -gain estimates is small. The dynamic event-generators even further reduce this conservatism, and as a result typically generate significantly fewer events than their static counterparts, while guaranteeing the same control performance. The benefits of these new event-generators are demonstrated via two numerical examples.
AB - In event-triggered control (ETC) systems, the measured state or output of the plant is sent to the controller at so-called event times. In many ETC systems, these event times are generated based on a static function of the current state or output measurement of the system and its sampled-and-held version that is available to the controller. Hence, the event-generator does not include any dynamics of its own. In contrast, dynamic event-generators trigger events based on additional dynamic variables, whose dynamics depend on the state or output of the system. In this paper, we propose new static and dynamic continuous event-generators (which require continuous measuring of the plant output) and periodic eventgenerators (which only require periodic sampling of the plant output) for linear control systems with communication delays. All event-generators we propose lead to closed-loop systems which are globally exponentially stable with a guaranteed decay rate, L2 -stable with a guaranteed L2 - gain, and have a guaranteed positive minimum inter-event time. By using new Riccati-based analysis tools tailored to linear systems, the conservatism in our decay rate and L2 -gain estimates is small. The dynamic event-generators even further reduce this conservatism, and as a result typically generate significantly fewer events than their static counterparts, while guaranteeing the same control performance. The benefits of these new event-generators are demonstrated via two numerical examples.
KW - Linear matrix inequalities
KW - Linear systems
KW - Networked control systems
KW - Riccati equations
UR - http://www.scopus.com/inward/record.url?scp=85048175817&partnerID=8YFLogxK
U2 - 10.1109/TAC.2017.2713047
DO - 10.1109/TAC.2017.2713047
M3 - Article
VL - 63
SP - 174
EP - 188
JO - IEEE Transactions on Automatic Control
JF - IEEE Transactions on Automatic Control
SN - 0018-9286
IS - 1
ER -