Event-triggered control for discrete-time systems

Abstract

Abstract-In this paper, event-triggered strategies for control of discrete-time systems are proposed and analyzed. Similarly to the continuous-time case, the plant is assumed input-tostate stable with respect to measurement errors and the control law is updated once a triggering condition involving the norm of a measurement error is violated. The results are also extended to a self-triggered formulation, where the next control updates are decided at the previous ones, thus relaxing the need for continuous monitoring of the measurement error. The overall framework is then used in a novel Model Predictive Control approach. The results are illustrated through simulated examples.

Periodic event-triggered control of nonlinear systems using overapproximation techniques

“…Note that the redesigned event functionΓ may become positive in between sampling instants along solutions to (6). However, whenΓ(z(t s j )) < 0 at sample time t s j , j ∈ N 0 , our design ofΓ given by (15) ensures that Γ(z(t)) < 0 will be satisfied for all t ∈ [t s j , t 0 for all t ∈ R >0 , and thus preserves the control performance guarantees of the original CETC system (3).…”

“…Note that PETC differs from standard periodic sampled-data control, as in PETC the event times are only a (specific) subset of the sampling times and can be aperiodic. Of course, event-triggered control schemes for discrete-time systems (e.g., [4,6,8,15,18,27]) can also be interpreted as PETC schemes, but these do not take into account the inter-sample behavior. In the past few years, various PETC strategies have been proposed, see, e.g., [9][10][11]13,21].…”

Output Feedback Event-Triggered Control

“…Periodic event-triggered control is another alternative in which the plant measurements are sampled periodically and, at each sampling instant, the event-triggering condition is evaluated to decide whether or not to transmit new measurements and control signals, see e.g. [7], [44], [27], [40], [42], [87]. The main benefit of this strategy is that the Zeno phenomenon is ensured to be avoided since the periodic sampling interval serves as a guaranteed lower bound on the inter-transmission times.…”

Reducing usage of the computational resources by event driven approach to model predictive control

“…This approach exploits predictive nature of some control methods (eg MPC) to include predicted control signals into otherwise unused parts of the network message, in other words, a sequence of future control signals is transmitted through the network at each update time instead of single control signal [7]. There are also works in existence, which combine the ideas of the ETC and packet-based predictive control [8,9,10]. In [8] an eventtriggered MPC setup for unconstrained systems was presented using input-to-state stability notions as basis, later extension of this work [9] proposed ETC schemes for constrained discrete time systems.…”

“…There are also works in existence, which combine the ideas of the ETC and packet-based predictive control [8,9,10]. In [8] an eventtriggered MPC setup for unconstrained systems was presented using input-to-state stability notions as basis, later extension of this work [9] proposed ETC schemes for constrained discrete time systems. Especially noteworthy is the fact that this extension utilizes optimal control sequences produced by MPC optimization problem in an open-loop manner between update times.…”