Facilitating reuse of control software through context modelling based on the six-variable model

Abstract

When control software is developed, the context of the software is not predefined or given. At first, certain properties of objects in the real world need to be monitored/controlled and developers decide which sensors, actuators, other systems to use to monitor/control them, i.e. they make contextual decisions. Frequently, it is not possible to monitor/ control exactly these real world properties. Instead, a different set of properties is monitored/controlled whose values are related to the real world properties. Existing approaches like the famous Four-Variable Model call for documenting the monitored, controlled, input, and output variables for a control software. Yet, they do not ask for documenting the properties that have been of interest at first, before deciding which sensors, actuators, other systems to use, i.e. the real world properties. This results in problems when the control software shall later on be reused in another context. The new context may, for example, comprise additional sensors. In this situation, it is hard for developers to decide which input variables are still necessary and should somehow be monitored and which ones not. To avoid such problems, we suggest a context modelling method which is based on our extension of the Four-Variable Model, the Six-Variable Model, and needs to be applied during requirements engineering.