Apparatus and method for controlling a paper machine or other machine using measurement predictions based on asynchronous sensor information

An apparatus, comprising:

• at least one memory operable to store a continuous time model representing a process controlled by a controller;
• and
• at least one processor operable to:
• generate multiple discrete time models of the process using the continuous time model;
• predict measurements or states to be used by the controller to control the process, the predicted measurements or states generated using the discrete time models;
• provide the predicted measurements or states to the controller such that the controller uses the predicted measurements or states at a sampling rate of the controller;
• and
• update at least some of the predicted measurements or states using measurements associated with a characteristic of an item from a sensor;
• wherein the discrete time models are associated with multiple sampling times, wherein different sampling times are associated with different types of events and each discrete time model being associated with at least one event not associated with any other discrete time model.

A computer readable medium embodying a computer program, the computer program comprising computer readable program code for:

• generating multiple discrete time models of a process to be controlled using a continuous time model of the process;
• predicting measurements or states to be used by a controller to control the process, the predicted measurements or states generated using the discrete time models of the process;
• providing the predicted measurements or states to the controller such that the controller uses the predicted measurements or states at a sampling rate of the controller;
• and
• updating at least some of the predicted measurements or states using measurements associated with a characteristic of an item from a sensor;
• wherein the discrete time models are associated with multiple sampling times, and wherein different sampling times are associated with different types of events and each discrete time model being associated with at least one event not associated with any other discrete time model.