Human-Machine Interaction - Views - Design - Introduction

    Attention : Script is not a complete representation of the oral lecture !!! 
Script is not yet completely finished !!!

AUTHOR: Gerd D´┐Żen-Henisch
EMAIL: doeben_at_fb2.fh-frankfurt.de

Design View


The design view is different compared to the observation view. The HMI-Experts are looking to the realworld problems not to 'measure' some 'properties' of human machine interactions but to 'model' those properties which are important for human machine interactions (look for the following to the whiteboard sketches below)

Figure: Whiteboard-figures related to the topic design view

The main question here is, what are those properties which are necessary for modeling?

Within the real world, the domain of possible applications is primarily characterized either as objects which are agents which can act or objects which are not agents. Furthermore can objects have different kinds of properties/ attributes.

An additional property of humanlike agents  is the hyothesis that they posses internally 'mental models'. These mental models can not directly be observed, but assuming these mental models is very useful in the explanation of the observable behaviour of these agents.

The mental models are functional effectiv for the interpretation of the perceptions, for the planning and for the concrete actions. Mental models are no static entities. They are highly dynamic: they evolve during time and they can fade away. The development depends from many factors. To some extend one can call this evolvement learning and forgetting.

Thus it makes sense to incorporate these concepts (objects, properties, agents, actions, mental model....) into that kind of modeling which  is characteristic for human machine interaction.

It can furthermore be assumed that all those aspects of the real world, which are important for the actions of an agent (e.g. relevant objects and their properties, other agents, possible actions ...), are  somehow represented in the mental model. If this is not the case,  then conflicts arise which challenge the agent to adapt his mental model. There is no guarantee that he can do this.

The main goal of an HMI-Expert is the construction of an user-interface (UI)  for a technical system which has to serve some users (H) such, that they can solve some defined tasks (T) with the aid of this interface. To distinguish between the 'given' user interface of a technical system and the 'consciously designed' user interface we will speak of 'designed user interfaces' (DUI) and the  'bare machine' (BM). The bare machine represents a  collection of possible interactions with a technical system (M) 'as it is'. The designed user interface is the result of a consciously designed process which establishes an artificial interface based on the poor interface.

 The designed user interface can be an audio interface, a graphical interface, a haptic interface etc. or a multimodal interface combining different modes of perceptions and actions.

But to know, which properties are needed for a  designed user interface, one has to model the tasks which a user should solve with the aid of this interface.  Thus after the completion of the  description of the requirements for some kind of an HMI-problem, the HMI-Expert has to model all the tasks which are induced by this requirements description.

Figure: Task graph

Modeling a task can be done in many different ways. One way is to look at a task as a set of possible sequences starting with a certain situation as the 'Beginning' and then generate possible sequences of situations by doing some actions. Every action does change a situation. Thus a task graph can be seen as a formal model of a task with at least one start situation and at least one goal situation. Tasks graphs can by acyclic or cyclic. The number of involved objects as well as the length of the necessary sequences can be a measure of the complexity of the task.

Additional one can model the necessary mental model of the human agent as part of such a situation.