Event-triggered model predictive control with machine learning for compensation of model uncertainties
Authors:
- 1 December 2017
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 5463-5468
Abstract
As one of the extensions of model predictive control (MPC), event-triggered MPC takes advantage of the reduction of control updates. However, approaches to event-triggered MPCs may be subject to frequent event-triggering instants in the presence of large disturbances. Motivated by this, this paper suggests an application of machine learning to this control method in order to learn a compensation model for disturbance attenuation. The suggested method improves both event-triggering policy efficiency and control accuracy compared to previous approaches to event-triggered MPCs. We employ the radial basis function (RBF) kernel based machine learning technique. By the universial approximation property of the RBF, which imposes an upper bound on the training error, we can present the stability analysis of the learning-aided control system. The proposed algorithm is evaluated by means of position control of a nonholonomic robot subject to state-dependent disturbances. Simulation results show that the developed method yields not only two times less event triggering instants, but also improved tracking performance.Keywords
This publication has 16 references indexed in Scilit:
- Deep learning helicopter dynamics models2015
- Distributed estimation using online semi‐supervised particle filter for mobile sensor networksIET Control Theory & Applications, 2015
- Reachability-based safe learning with Gaussian processes2014
- A nonlinear quadrotor trajectory tracking controller with disturbance rejectionControl Engineering Practice, 2014
- Convergence of a PI coordination protocol in networks with switching topology and quantized measurements2012
- Event-Triggered Strategies for Decentralized Model Predictive ControllersIFAC Proceedings Volumes, 2011
- Input-to-state stable MPC for constrained discrete-time nonlinear systems with bounded additive uncertainties2004
- A characterization of integral input-to-state stabilityIEEE Transactions on Automatic Control, 2000
- Support-vector networksMachine Learning, 1995
- Universal Approximation Using Radial-Basis-Function NetworksNeural Computation, 1991