Context-aware systems for assisting users in their activities of daily living : from activity recognition to augmented reality assistance provision

智能城市和智能家居及其传感器和设备是很好的技术环境，可以在其中部署上下文感知系统，并可以收集大量信息以支持活动识别和上下文感知服务提供。我们的第一个目标（Obj.1）是设计策略和算法，以便使用混合分类和模式识别算法以及贝叶斯模型来识别和预测人类活动。这种算法必须管理数据的异构性，考虑硬件限制（例如电池寿命），并考虑对移动设备上的数据进行准实时分析。在我们最近的工作中，我们开发了基于混合系统的策略和算法来分析上下文信息，以识别城市环境中移动用户的活动。围绕Obj.1的研究活动将是通过增加识别活动的范围来追求城市环境中移动用户的在线活动识别，并将更多的上下文信息（如周围用户或时间逻辑）纳入识别过程。

Smart cities and smart homes, with their sensors and devices, are great technological environments in which context-aware systems can be deployed and where a quantities of information can be collected to support activity recognition and context-aware service provision. Our first objective (Obj.1) is to design strategies and algorithms in order to recognize and predict human activities with hybrid-classification and pattern recognition algorithms as well as Bayesian models. Such algorithms must manage heterogeneity of data, take into account the hardware limits (e.g. battery life), and consider a quasi-real-time analysis of the data on mobile devices. In our recent work, we have developed strategies and algorithms to analyze contextual information to recognize activities of mobile users in urban environment based on hybrid systems. The research activities around the Obj.1 will be to pursue the online activity recognition of mobile users in urban environments by increasing the range of recognized activities and include more contextual information into the recognition process such as surrounding users or temporal logic. If context-aware applications can provide value-added services to general users, such software can provide extended assistance to dependent users (e.g. seniors), for instance, by helping them in the realization of activities of daily living. Our second objective (Obj. 2) focuses on interdisciplinary work in ubiquitous computing and human-computer interaction, which consists of the conception and development of context-aware services that assist users in the realization of ADLs. More specifically, we are working on two sub-objectives : (a) the design and evaluation of augmented reality cognitive orthoses to support people in the realization of ADLs; (b) design and evaluation of multimodal interactions techniques in AR in order to enhance usability of cognitive orthoses content provided in Obj. 2.a and take into account the limitation of users (e.g. physical or task related limitation). Providing tasks and cognitive assistance in AR includes several challenges from the human factors and technological perspective, such as 3D orientation and interaction, multimodal interactions, attention and mental charge, etc. The results of my research will contribute significantly to the state of the art of context-awareness, ambient intelligence and user assistance research fields. Our research program will also contribute to the formation of the highly qualified personnel (HQP) involved in our research by providing a high-quality academic education and a dynamic research environment. Finally, we are also working toward having a positive impact on the Canadian population in general by transferring the developed theories, models, and technologies to public and private organizations, but also toward people with special needs by increasing their autonomy and quality of life.