Personalization platform for multimodal ubiquitous computing applications

Abstract

We currently live surrounded by a myriad of computing devices running multiple applications. In general, the user experience on each of those scenarios is not adapted to each user’s specific needs, without personalization and integration across scenarios. Moreover, developers usually do not have the right tools to handle that in a standard and generic way. As such, a personalization platform may provide those tools. This kind of platform should be readily available to be used by any developer. Therefore, it must be developed to be available over the Internet. With the advances in IT infrastructure, it is now possible to develop reliable and scalable services running on abstract and virtualized platforms. Those are some of the advantages of cloud computing, which offers a model of utility computing where customers are able to dynamically allocate the resources they need and are charged accordingly. This work focuses on the creation of a cloud-based personalization platform built on a previously developed generic user modeling framework. It provides user profiling and context-awareness tools to third-party developers. A public display-based application was also developed. It provides useful information to students, teachers and others in a university campus as they are detected by Bluetooth scanning. It uses the personalization platform as the basis to select the most relevant information in each situation, while a mobile application was developed to be used as an input mechanism. A user study was conducted to assess the usefulness of the application and to validate some design choices. The results were mostly positive.