Output feedback stochastic MPC with packet losses

Yan, Shuhao; Cannon, Mark; Goulart, Paul J.

doi:10.1016/j.ifacol.2020.12.506

Cited by 3 publications

(2 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Bertsekas, 1995, Van Parys et al, 2013, Kouvaritakis et al, 2003, Kamgarpour and Summers, 2017, Feinberg and Shwartz, 1999, and an appropriate discounting factor can provide stability guarantees [Postoyan et al, 2017]. Relative to our earlier work in Yan et al [2020], this paper includes additional results on discounting and stability and more extensive numerical studies. The main contributions are as follows:…”

Section: Introductionmentioning

confidence: 99%

Stochastic output feedback MPC with intermittent observations

Yan,

Cannon,

Goulart

2020

Preprint

Self Cite

View full text Add to dashboard Cite

This paper considers constrained linear systems with stochastic additive disturbances and noisy measurements transmitted over a lossy communication channel. We propose a model predictive control (MPC) law that minimises a discounted cost subject to a discounted expectation constraint. Sensor data is assumed to be lost with known probability, and data losses are accounted for by expressing the predicted control policy as an affine function of future observations, which results in a convex optimal control problem. An online constraint-tightening technique ensures recursive feasibility of the online optimisation problem and satisfaction of the expectation constraint without imposing bounds on the distributions of the noise and disturbance inputs. The discounted cost evaluated along trajectories of the closed loop system is shown to be bounded by the initial optimal predicted cost. We also provide conditions under which the averaged undiscounted closed loop cost accumulated over an infinite horizon is bounded. Numerical simulations are described to illustrate these results.

show abstract

Section: Introductionmentioning

confidence: 99%

Stochastic output feedback MPC with intermittent observations

Yan,

Cannon,

Goulart

2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…The application of optimal control to stochastic model has been extensively used for Stochastic Optimal Controllers and particularly in Stochastic Receding Horizon Control algorithms, also widely known as Stochastic Model Predictive Control (SMPC) [12], [13]. [14] applied SMPC to models with communication impairments, but in that case, the optimal trajectory is taken as determined and used as the origin of the states space at each time step: the state then is the deviation from the reference and has to be minimized (there is not an optimal trajectory computation). The problem of lossy networks and system uncertainty on MPC are considered separately in [15] and [16], respectively.…”

Section: Introductionmentioning

confidence: 99%