In this section, we select frameworks showing spatial reconfigurability and having more relevance to ambient environments among different domains such as ambient intelligence (Aarts, 2004), ubiquitous computing (Weiser, 1991), and responsive architecture (Bullivant, 2005; Sterk, 2003). Conceptual, representational, and computational frameworks are reviewed. We describe each as follows.
2.2.1 Hybridized Control Model
(Sterk, 2003) proposed an extensible model for controlling responsive architecture, which can be used as a fundamental concept to describe reconfigurable ambient environments. It is a simple model which consists of three parts (Fig 2-2):
1) User input: which offers users the means to control and interact with the building;
2) A building structure: which has a responsive capability that enables it to directly respond to environmental loads, and
3) Spatial responses: which is used to control the partitioning or services for activity inside spaces.
Fig 2-2: A hybridized control model for responsive architecture (after Sterk, 2003).
2.2.2 Frameworks of Context Awareness
In general, the context is the circumstance of “who is in where, doing what, how, and for what”. In an ambient intelligence (Marzano and Aarts, 2003) viewpoint, objects and services need to be aware of the state of their surroundings at any given moment. It is a fundamental technology used to achieve a user-centric intelligent
environment. There are many ways to describe or structure contexts. For example, (Oh and Woo, 2004) proposed a unified model to format and integrate contexts into the structure of 5W1H (Why, Who, When, Where, Which, and How).
There are three phases in the workflow cycle of context awareness (Aarts and Roovers, 2003):
1) Perceiving the environment: the first step is to collect information about the environment and turn it into a useful form with sensor technology or smart sensors.
2) Classifying and analyzing the data: the second step is to use the information provided by sensors to determine the state of the environment as a whole based on the model of context.
3) Interpreting the context and taking action: The last step is, based on the environmental context the system has perceived, to use high-level knowledge to decide what the system should do.
2.2.3 Implicit HCI Framework
iHCI (Implicit Human-Computer Interaction) is a conceptual model which takes
context into account as implicit input and has influence on the environment by implicit output (Fig 2-3) (Schmidt, 2000). It is most suitable for systems where the the user should not be distracted from the main task in the physical spatial context.This model is widely applicable for specific domains, such as proactive applications, adaptive UIs, user interruptions, communication applications, resource management, and the generation of metadata.
(Schmidt, 2004) gave a formal definition of implicit input and implicit output:
Implicit Input: Implicit input pertains to the actions and behaviors of humans,
which are done to achieve a goal and are not primarily regarded as interaction with a computer, but captured, recognized, and interpreted by a computer system as input.Implicit Output: Output of a computer that is not directly related to an explicit
input and which is seamlessly integrated with the environment and the task of the user.Fig 2-3: Implicit human-computer interaction model (after Schmidt, 2004).
2.2.4 Workflow Cycle of Ambient Intelligence
(Hellenschmidt and Wichert, 2005) proposed three major steps for the workflow cycle of an ambient intelligence system:
Awareness: The environment and the objects within the environment should be
aware of the user’s current situation, his interaction condition with the environment, his personal condition, and the possible condition he should be adapted to.Intention Analysis: The environment must infer the user’s intention based on the
situation it is aware of, and respond with possible cooperative or proactive support to the user.Strategy Planning and Execution: The environment should transform the user
intention it inferred into an adaptation strategy which the environment and environmental objects can provide.Such an interaction cycle can be generalized and termed as
‘Goal-based Interaction’ (Heider and Kirste, 2002), as shown in Fig 2-4. Goal-based interaction requires two functionalities: Intention Analysis, interpreting user interactions and environmental contexts into concrete goals, and Strategy Planning, which maps goals to device operations (Hellenschmidt and Wichert, 2005).
Fig 2-4: Principle of goal-based interaction (after Heider & Kirste, 2005).
2.2.5 Framework for Project Aura
The pervasiveness of computers frees users from being bound to specific desktop computers. Based on such a concept, project Aura (Sousa and Garlan, 2002) tries to create a distraction-free computing environment with ubiquitously available computational resources. Users in a ubiquitous computing environment can bind and compose their own task context and release them at any physical service hot-spot. The computer-supported task thus becomes a personal aura which surrounds people, provides personalized settings, and requires no configuration efforts for switching among different computational platforms or environments.
2.2.6 SODAPOP
SODAPOP (Self-Organizing Data-flow Architectures supPorting Ontology-based problem decomPosition), is a middleware for simplifying the framework integration process among smart objects (Hellenschmidt and Kirste, 2004). Its objective is to make the environment observe and analyze the user’s goal, and combines appropriate components with environmental resources into a unifying system automatically in real-time according to the analyzed data. The integration has two aspects: 1) Components integration: the pattern matching in the system level. For example, attaching an input device to the ensemble’s interaction event bus; 2) Operational Integration: The mapping of interface operations with metaphorical relation. For example, connecting a CD player into a CD recorder could be embodied as “Copy” (Hellenschmidt and Kirste, 2004).
2.2.7 Framework for VICOM and PER2
Inspired by Antonio Damasio’s Human Conscience Model (Damasio, 2000), (Marchesotti et al., 2005) designed a flexible architecture the for user to interact with ambient intelligence environments (Fig 2-5). This model is comprised of a few mapped parts: a) Eso-Sensors for sensing external contextual data, b) Endo-Sensors for sensing internal status, and c) Self Kernel, which connect to Autobiographical
Memory (i.e. short-term memory) and Autobiographical Self (i.e. long-term
memory). This neurobiologically-inspired model is the basis of the artificial analysis and decision core, allowing the system to acquire and manage a deeper understanding of context information. The flexibility of the architecture is issued inthree aspects: context-awareness, multimodal communication, and user-centered adaptive interaction. The proposed design employs a rule-based adaptation module where acquired contextual knowledge about the environment and the user is represented in terms of concepts and facts and are exploited to personalize the multimodal feedback for the user (Stefano et al., 2005).
Fig 2-5: Framework for VICOM and PER2 (after Stefano et al., 2005)
2.2.8 Summary
In this section we have reviewed representational as well as computational frameworks which could be generally applied to describe a triggering mechanism with sufficient reconfigurability.
Though not developed specifically for ambient reconfiguration purpose, each framework has provided us with insights for developing an ambient reconfigurable framework. Mainly, the Hybridized Control Model (Sterk, 2003) has framed a general and fundamental framework for approaching ambient reconfiguration.
Frameworks of Context Awareness (Aarts and Roovers, 2003; Marzano and Aarts, 2003; Oh and Woo, 2004) have illustrated the computational workflow mapping to spatial contexts. Implicit HCI frameworks (Schmidt, 2000) have suggested implicit interaction as a less distractive way for the user to interact with a physical system.
The workflow cycle of ambient intelligence (Heider and Kirste, 2002;
Hellenschmidt and Wichert, 2005) have suggested a goal-based type of interaction, which point out that the need for ambient interactive services is based on the user’s intention. Aura (Sousa and Garlan, 2002), a work that approaches distraction-free interaction by providing users with ubiquitously accessible platforms, has suggested automatic configuration of spaces without the user’s explicit configuring process. SODAPOP (Hellenschmidt and Kirste, 2004; Hellenschmidt and Kirste, 2004) suggested that different combinations of operations or objects refer to different interactive contexts, and the framework of VICOM and PER2 (Marchesotti et al., 2005) (Stefano et al., 2005) have provided us with a flexible and intelligent model that is suitable for achieving ambient reconfiguration technologically. Some processing techniques in terms of contextual data acquisition and categorizations were also learned.
To gain a whole view on the research problem and to understand the practicability as well as feasibility in terms of ambient reconfiguration, the next section reviews related projects that evoke ambient reconfiguration from an interface perspective.