A Real-time Task Model Based on Ideal Instant
Fábio Rodrigues de la Rocha, Rômulo Silva de Oliveira · 2006
In many digital control applications, data acquisition and process control are time-critical actions, assumed to be instantaneous and strictly periodic. However, aspects related to real-time scheduling algorithms may postpone the task execution and, consequently, delay the data acquisition and process control. We consider the acquired data from sensors as valid if acquired inside a time-interval that surrounds an ideal instant which is the point of maximum system's benefit. The data acquired before or after this time-interval are considered useless to the controller. We propose a new task model to deal with the variation in the execution time of a specific point inside tasks. In this model, tasks have an ideal instant to present their results or acquire data from sensors (QoS metric)