Mobile Intelligence

20.9 CONCLUSION AND FUTURE WORK

In this chapter, we presented VOICE, a framework for developing mobile context-aware systems with speech interface accessible over a telephone connection. To prove the feasibility of the architecture of VOICE, we implemented a Restaurant Search Guide, a simple context-aware recommender system that helps its users with finding a restaurant according to multiple combinations of criteria; the system connects to multiple data source providers to obtain restaurant data and to generate driving directions to the chosen restaurant. We plan to further validate this solution by conducting a usability study based on the analysis of the system usage patterns.

There are several directions for adding functionality to VOICE architecture. Mixed-initiative dialogs could be used to simplify the number of prompts presented to the user in order to obtain a minimally necessary amount of data needed for a successful execution of queries. To streamline the process of customization of VOICE during the development of specific applications, special templates could be developed for different types of Data Source Adapters corresponding to a local database, an XML Web Service, or an application-specific or platform-specific API. Furthermore, it could be possible to mix some portions of the users' voice-based interface with SMS-based interface; for example, the users might want to query the system using the voice interface but receive the results as a text message on their mobile ...