Reference:
P.J. van Overloop,
J.M. Maestre,
A.D. Sadowska,
E.F. Camacho, and
B. De Schutter,
"Human-in-the-loop model predictive control of an irrigation canal,"
IEEE Control Systems Magazine, vol. 35, no. 4, pp. 19-29,
Aug. 2015.
Abstract:
This article contributes towards extending the scope of
human-in-the-loop (HIL) control for systems when human operators serve
as actuators or sensors. In particular, this work concerns those
situations where humans are operators of the control system requiring
their actions on a regular basis. That is, no human decision is
involved in the control, although the control system relies on the
operators to implement the control actions and to perform
measurements.
In this article, a large-scale system consisting of cascade-connected
subsystems that influence each other through mutual interrelations is
considered. Although there might be local automatic controllers within
each subsystem, the actions of the human operators form the nucleus of
the overall system operation. More specifically, it is assumed that
there are a number of operators working within the system as sensors
and actuators. However, the fact that the number of operators is less
than the number of subsystems in the large-scale system directly
implies that both the sensing and the actuating processes have a
sparse nature, which diminishes the performance of the system with
respect to standard fully-automatic methods.
The key idea of the HIL approach presented in this article is to
optimize the operators’ work in real time by integrating their labor
into an online optimization problem that maximizes the performance of
the system. In addition to operating in real time, it is also
convenient to explicitly consider event-triggered approaches. The
contributions of this article are twofold. Primarily, a novel HIL-MPC
scheme for a large-scale system with multiple operators to serve as
sensors and actuators is presented. Given the mobility of the
operator, the new approach is thereafter referred to as Mobile MPC
(MoMPC). Secondly, the MoMPC approach is tested on an accurate
numerical model of an irrigation canal, namely the Dez canal in Iran.
In this way, a realistic performance evaluation of MoMPC can be
executed.