DEPARTMENT OF COMPUTER SCIENCE UNIVERSITY OF WATERLOO SEMINAR ACTIVITIES MASTER'S THESIS PRESENTATION -Monday, July 21, 1997 Ryan Gunther, graduate student, Dept. Comp. Sci., Univ. Waterloo, will speak on ``The Use of 3D Sound as a Navigational Aid in Virtual Environments''. TIME: 10:30-11:30 a.m. ROOM: DC 1304 ABSTRACT Because virtual environments are less visually rich than real-world environments, careful consideration must be given to their design, otherwise much of the potential for virtual reality will be lost. One important design criteria is to make certain that adequate navigational cues are incorporated into a virtual environment. The usability of virtual environments will be decreased if users often become lost or disoriented, as is frequently the case now. This thesis investigates the use of spatialized sound as a navigational aid for people using virtual reality systems. An experiment was performed to determine if incorporating spatialized sound into a virtual environment a) helps people find specific locations in the environment and b) influences the rate at which a person acquires configurational knowledge of the environment. The results show that the addition of spatialized sound to a virtual environment does improve the ability of people to find their way around an environment, although the addition of sound does not increase the amount of configurational knowledge a person is able to acquire. In fact, it appears that the addition of sound may suppress the development of configurational knowledge.
By applying database technologies to embedded control applications such as telephony software, a control program can be made physically data independent by specifying access and update of control data in a declarative query language such as OQL. In such a memory-resident and real-time environment, the generated access plans must avoid the use of temporary storage for arbitrarily sized intermediate results and the use of expensive operations such as sorting. Also, to enable increment use on legacy control software, the generated plans must be able to navigate existing data structures.
This thesis presents an optimization approach for such applications by using existential query graphs (EQG)-a refinement of Peirce's existential graphs. An EQG is expressive enough to directly capture an extended class of conjunctive queries, as well as access plans that encode strategies for query evaluation.
To simplify the mapping from an OQL query to an EQG, a canonical form has been introduced. The mapping algorithms are then discussed. A join order selection procedure that incrementally refines a pair of EQGs, namely, a query graph and a plan graph, is also presented. The former graph captures what is necessarily true of the query, while the latter encodes an access plan for the query. Differential analysis of the two graphs distinguish query conditions that require special checks in access plans from constraints that are defined in the schema.
The expansion of the new features in telephony service has uncovered many unsuspected problems. This thesis addresses on detection feature interactions at the requirement phase. A feature provides added functionality to an existing service in the telecommunication system. Feature interaction occurs when the behavior of an existing feature is changed by the presence of a new feature.
This thesis presents an approach for detecting feature interactions by assertion analysis. Assertions are property facts and assumptions about the telephony system that the feature expects to hold. Each telephony service and feature is modeled by a state-transition- machine in our approach. A formal tabular notation is used to specify services and features individually. Specifications of feature assertions are developed in terms of sets and relations. Assertion analysis is based on composing the feature specifications automatically during reachability analysis, and interactions are detected at reachable states during the composition. Evaluation are presented with respect to a benchmark paper published by the Bellcore.