|
||||
|
Cloud2Bubble: a context-aware platform for enhancing quality of experience
| Permalink In recent years, important technological developments, such as hardware component miniaturization and modern broadband networks, have paved the way for the mass adoption of mobile devices and ubiquitous communication networks. This trend has largely contributed to a shift in the human-computer interaction (HCI) paradigm. The personal computer is giving way to a multiplicity of interconnected devices. These devices, supported by a powerful cloud computing infrastructure, are being transformed into portals to a virtual space, composed of users' data and applications that are aware of users' preferences, needs, activities, and even their surrounding environment. In this context, in which a pervasive computing environment is becoming a reality, both technical and cognitive issues arise.1 On a technical level, it is a complex task to integrate and manage a number of heterogeneous devices with different capabilities, while different applications themselves require different levels of data quality. From the users' perspective, their focus is shared between other activities in their environment and devices with limited capabilities, such as screen size and connectivity. Finally, the increasing number of services available, made possible by this same environment, offers users a sometimes bewildering number of options when searching for relevant applications. Human cognitive capacity limits users' ability to deal with this overwhelming environment, in effect creating a virtual ‘bubble’ composed of their known applications and services. Therefore, to enrich the user experience (UX), users need to be closely integrated within the loop of interaction. This integration enables a system to be aware of the quality of its services and how to act accordingly. Context-aware systems are capable of collecting different sources of data and constructing a context model that characterizes a given entity: a person, place or object, including the user and the applications.2 Reasoning based on this context model enables the system to make decisions that have a potentially positive impact on a user's quality of experience (QoE); in other words, the degree to which a system meets users' expectations of their experience.3 Cloud2Bubble (C2B), depicted in Figure 1, addresses these issues by defining a model that aggregates environment-generated data, including affective data, from the system (the cloud) and delivering relevant services to the end user (the bubble). The C2B model is a combination of an expandable domain service and complex event processing, based on a rule engine and fuzzy logic.4 This model collects environment data using widely deployed technology, ranging from environment sensors to mobile phones. The model relies on intelligent policy-based computing5 to dynamically adjust its service delivery to users' preferences and expectations. Figure 1. Cloud2Bubble aggregates environment-generated data, including affective data, from the system (the cloud) and delivers relevant services to the end user (the bubble). 
We focus on the UX aspect of such a system with the goal of enhancing QoE. To do so, we introduce emotion as a dimension of interaction, a concept from the field of affective computing.6 This concept relates the characteristics of the surrounding environment to users' affective responses, building a profile that describes whether the environmental attributes have a positive or negative impact on users' emotional states (see Figure 2). This affective profile is then used either to improve the environment, e.g., by adjusting the temperature in a room, or to suggest other, more suitable alternatives. Figure 2. Model of Cloud2Bubble operating in a public transportation environment. 
However, data collection is a sensitive subject, and the possibility of describing user behaviour to such a level of detail raises serious security and privacy concerns. Therefore, such systems must take these concerns into account and ensure they are designed to maintain user anonymity and to be able to prevent malicious access to data and functions. A model based on intelligent policy-based computing provides such a platform, capable of flexibly dealing with the complexity of sensitive data exchange, while protecting and maintaining the confidentiality of the data. From an HCI perspective, applications need to be adaptive, while maintaining intelligibility. A decision to adapt part of an application or the environment based on insufficient or inaccurate information might degrade the UX rather than improve it. This will have the opposite of the desired effect on QoE. In addition, disruptive or unclear adaptation, even if based on correct assumptions, might not be well received. Therefore, the integration of such features into users' activities must follow a systemic and seamless approach. The C2B concept is potentially applicable to a wide spectrum of socio-technical environments, from healthcare to retailing. The context of personal mobility is especially attractive as a result of its dynamic and technical nature. For example, modern public transportation networks already provide contextual data, to a degree. For instance, Transport for London has made detailed information public, updated every minute. In addition, there is a need for intelligent traveller information systems that can provide travellers with relevant information about their journeys. This would include price and duration information, but could also, for example, encompass information on how convenient and comfortable the journey might be.7 In summary, we believe modern pervasive environments expose a wide range of possibilities, but they also present several challenges. Cloud2Bubble addresses these challenges by proposing a model for collecting environment-generated data and delivering a service tailored to users according to their preferences, needs and affective state. The context of public transportation is particularly attractive for such systems because of its technical development in modern networks. In the future, we plan to investigate certain issues raised by the possibility of intelligent adaptive applications. For instance, how would users respond to incentives to adapt their behaviour? How would such a system affect the relationship between macro-level goals (at a population level) and micro-level behaviour (at the individual level)? In particular, would making people more aware of their impact on the environment support a positive and sustainable change in behaviour? References
Stay Informed
|
||||
|
||||