vAssist : voice based AAL services
Résumé
Currently, 87 million persons in the EU-27 are aged 65+. With increasing age, the prevalence of physical and cognitive problems increases, resulting in a high demand for appropriable services to avoid a technological divergence, a separation which is not only driven by people's social or economic background but also related to an increasingly aging society. These facts not only support the claim for advanced services but also for interaction paradigms that enable the compensation of age-related restrictions. In direct relation, the vAssist project focuses on supporting seniors suffering from fine-motor problems or seniors with little or no computer literacy by applying multi-language natural speech interaction. Studies indicate that speech interaction shows high acceptance rates among seniors. A reduction of costs is expected by achieving channel independence in the delivery of vAssist services so that existing hardware in the homes of seniors can be used. The use of devices users are already familiar with by just one interaction technique is expected to facilitate the access to the implemented and future services. The results presented concentrate on behaviors and requirements of seniors towards the features, the interaction and the delivery of the vAssist communication and well-being services. In addition first services together with results from an early evaluation study are presented. A User-Centered Market-Oriented Design process (UCMOD) considering user, technical and economic constraints is applied to ensure that user needs are properly addressed throughout the entire project. Focus groups combined with a Wizard of OZ simulation took place in Austria, France and Italy to gain information about the requirements regarding devices, services, interaction needs and business models. Results show that spoken input is perceived as a good alternative input technique especially for mobile devices; natural input language is preferred over commands and the overall engagement with the device is increased. The participants' preference for a "friend-like" interaction style and demand of giving the system a name could further indicate increased acceptance and trust. Therefore, VUIs can be seen as an efficient and promising input modality in particular if users suffer from age-related physical restrictions. Nevertheless, technical problems and insufficient component quality may reduce this potential. Hence, the implementation of error recovery strategies, feedback and an additional fallback modality is crucial. To sum up, vAssist holds great potential support accessibility and technology acceptance for older adults by implementing natural speech interaction and channel independence