Ubiquitous computing (Ubicomp) technologies are providing exciting new opportunities for enhancing physical spaces to support the needs and activities of people within them. The ability to develop such systems effectively will offer significant competitive advantage. Tools are required that will predict problems relating to use early in the design cycle. This proposal brings together a team of researchers and developers aimed at prototyping and calibrating methods and tools that facilitate the development of technology enhanced environments.
Ubiquitous computing systems pose new usability challenges that cut across all phases of design and development. We are particularly interested in “spaces” enhanced with sensors, public displays and personal devices. In evaluating and exploring these spaces it is not only necessary to explore conventional properties of usability but to explore properties of the environment that contribute to the experience of its users. Because of the potential cost of development and design failure, the characteristics of such systems must be explored using early versions of the system that could disrupt if used in the target environment. Being able to prototype and evaluate these systems early in the process is crucial to their successful development.
Similar work, e.g. UbiWorld , has a different focus in two important respects: (1) the proposed tool drives virtual environment simulations using models that can also be used for analysis; (2) context is a primary consideration in our models with a focus on how context affects the usability of the implicit interactions that are characteristic of these systems. No other research or development is addressing this problem in the context of a development process.
The PI has been researching the rigorous analysis of models of interactive systems [2, 3, 4] (c.f. the IVY workbench - POSC/EIA/56646/2004). More recent research has concerned extending models and analysis to an ambient intelligence setting. Models vary from qualitative models concerned with who gets what information when [5, 6] to quantitative models concerned with the behaviour of many users of technologically enhanced environments . How users experience such systems is particularly difficult to capture analytically, and yet is an important dimension of concern. We propose to develop a model-based prototyping environment that will allow for a more empirical style of analysis.
3D Application Servers such as SecondLifeTM or OpenSimulator provide a fast track to developing virtual worlds. They seem a natural choice for developing the type of prototype we are aiming at. Preliminary explorations carried out by the proposers  indicate the potential of the approach. OpenSimulator, in particular, has the advantage of being open source. This means the backend can be programmed, making it highly configurable and extensible.
The goals of the project are to: 1) employ an existing 3D Application Server to construct a rapid prototyping platform of ubiquitous computing environments driven by a model of the environment; 2) develop frameworks for modelling these environments with a particular focus on context; 3) explore the fidelity of these prototypes as a factor in evaluating the eventual human (usability and social) impact of the design.
To fulfil its goals the project will develop a software framework enabling rapid prototyping of ubiquitous systems targeted at built environments. The framework will plug into virtual reality (VR) simulations of the environments, and allow for the integration of users' portable devices thus immersing users within the environment using a mixture of reality and virtuality. Developers and users will be able to navigate the simulation to develop an impression of what it will be like to use the final system once fielded. Different levels of fidelity will be explored, from desktop VR to a CAVE .
A criterion for the framework’s success will be its fidelity as an evaluation tool. Comparative analysis of the user experience on an implemented benchmark system, and on different prototypes will enable this evaluation. Evaluation approaches will be researched to determine how well a given prototype fairs in relation to experience requirements. Drivers that determine this fitness (e.g. user stories or experience snapshots) must be identified. It is envisaged that user-stories will not be sufficient. Techniques are required to complement existing user-stories that enable developers to explore these more general properties of systems. These could, for example, include the animation of task models, or potentially problematic behaviours identified by verification analysis .
Another question relates to the expressiveness of the modelling notation adopted. The modelling of the benchmark system, together with interaction with its developers, will be the basis for this analysis.