This thesis studies the application of speech recognition and synthesis technologies, agents and virtual environments in the design of a digital assistant, helping a user of web-based e-commerce applications. A useful component of the user interface in an electronic commerce system is a speech user interface. This interface will contain a user's agent and serve an active role in recognizing a user's speech input as well as synthesizing the required speech sentences needed to be conveyed. The Speech GUI implemented in this thesis complements the web browser during virtual world e-commerce. The virtual world being researched in the MCRLab is a shopping mall, where users can enter a store and find sales agents ready to assist them. Each person within this virtual environment is represented with the use of an agent and avatar. The virtual environment is a multi-agent system. These agents (also represented by avatars) can navigate through the virtual environment such as a shopping mall, and interact with each other. Once dialog has occurred between a user's agent and the store's sales-agent and an item of interest is found, this item will be displayed in the virtual mall. The virtual environments technology, with the help of avatars will allow us to have good voice QoS coupled with synchronized avatar lip movements. The ultimate purpose of the Speech GUI is to assist a user in finding items of interest within a virtual shopping store. Also, when the items of interest are found, this Speech GUI will ask the web browser to display the objects within the virtual store. Communication for this is achieved using the HLA/RTI available standard. The second type of communication associated within the Speech GUI uses the KQML protocol to communicate with other peoples' agents.
| Identifer | oai:union.ndltd.org:uottawa.ca/oai:ruor.uottawa.ca:10393/8640 |
| Date | January 1999 |
| Creators | Hage, Ramsey. |
| Contributors | Georganas, Nicolas G., |
| Publisher | University of Ottawa (Canada) |
| Source Sets | Université d’Ottawa |
| Detected Language | English |
| Type | Thesis |
| Format | 130 p. |
Page generated in 0.009 seconds