Publicação
Design and evaluation of adaptive multimoldal systems
| Resumo: | This thesis focuses on the design and evaluation of adaptive multi-modal systems. The design of such systems is approached from an integrated perspective, with the goal of obtaining a solution where aspects related to both adaptive and multimodal systems are considered. The result is FAME, a model based framework for the design and development of adaptive multimodal systems, where adaptive capabilities impact directly over the process of multimodal fusion and fission operations. FAME over views the design of systems capable of adapting to a diversified context, including variations in users,execution platform, and environment. FAME represents an evolution from previous frameworks by incorporating aspects specific to multimodal interfaces directly in the development of an adaptive platform. One of FAME's components is the Behavioral Matrix, a multi purpose instrument, used during the design phase to represent the adaptation rules. In addition, the Behavioral Matrix is also the component responsible for bridging the gap between design and evaluation stages. Departing from an analogy between transitionnet works for representing interaction with a system, and behavioral spaces, the Behavioral Matrix makes possible the application of behavioral complexity metrics to general adaptive systems. Moreover,this evaluation is possible during the design stages,which translates into a reduction of there sources required for evaluation of adaptive systems.The Behavior al Matrix allows a designer to emulate the behavior of anon-adaptiveversionoftheadaptivesystem,allowing for comparison of the versions, one of the most used approaches to adaptive systems evaluation. In addition, the designer may also emulate the behavior of different user profiles and compare their complexity measures. The feasibility of FAME was demonstrated with the development of an adaptive multimodal Digital Book Player. The process was successful, as demonstrated by usability evaluations. Besides these evaluations, behavioral complexity metrics, computed in accordance with the proposed methodology, were able to discern between adaptive and non-adaptive versions of the player. When applied to user profiles of different perceived complexity, the metrics were also able to detect the different interaction complexity. |
|---|---|
| Autores principais: | Duarte, Carlos |
| Assunto: | Engenharia informática Teses de doutoramento |
| Ano: | 2007 |
| País: | Portugal |
| Tipo de documento: | tese de doutoramento |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório da Universidade de Lisboa |
| Resumo: | This thesis focuses on the design and evaluation of adaptive multi-modal systems. The design of such systems is approached from an integrated perspective, with the goal of obtaining a solution where aspects related to both adaptive and multimodal systems are considered. The result is FAME, a model based framework for the design and development of adaptive multimodal systems, where adaptive capabilities impact directly over the process of multimodal fusion and fission operations. FAME over views the design of systems capable of adapting to a diversified context, including variations in users,execution platform, and environment. FAME represents an evolution from previous frameworks by incorporating aspects specific to multimodal interfaces directly in the development of an adaptive platform. One of FAME's components is the Behavioral Matrix, a multi purpose instrument, used during the design phase to represent the adaptation rules. In addition, the Behavioral Matrix is also the component responsible for bridging the gap between design and evaluation stages. Departing from an analogy between transitionnet works for representing interaction with a system, and behavioral spaces, the Behavioral Matrix makes possible the application of behavioral complexity metrics to general adaptive systems. Moreover,this evaluation is possible during the design stages,which translates into a reduction of there sources required for evaluation of adaptive systems.The Behavior al Matrix allows a designer to emulate the behavior of anon-adaptiveversionoftheadaptivesystem,allowing for comparison of the versions, one of the most used approaches to adaptive systems evaluation. In addition, the designer may also emulate the behavior of different user profiles and compare their complexity measures. The feasibility of FAME was demonstrated with the development of an adaptive multimodal Digital Book Player. The process was successful, as demonstrated by usability evaluations. Besides these evaluations, behavioral complexity metrics, computed in accordance with the proposed methodology, were able to discern between adaptive and non-adaptive versions of the player. When applied to user profiles of different perceived complexity, the metrics were also able to detect the different interaction complexity. |
|---|