Towards Ambient Affective Intelligence and Interaction in Smart Environments

点击翻译按钮获取中文摘要

This program proposes research on the novel concept of ambient affective intelligence and interaction in smart environments (e.g. smart homes). The paradigm is at the cross-section of affective computing and ambient intelligence. Affective computing refers to modeling and recognition of human emotions while ambient intelligence is a paradigm that focuses on seamless integration of sensors and processors into everyday devices from a hardware perspective, and focuses on human-centricity, responsiveness, adaptiveness, personalization, and anticipation from an algorithm and interaction perspective. The overall goal of the program is to bring these two concepts together and create a new paradigm for smart environments and their elements to be pervasively and continuously aware of affective states of users, and react/interact accordingly.In order to achieve this objective, we will utilize a variety of wearable and Internet of things (IoT) sensing and actuation systems. We will first develop robust algorithms for continuous and real-time multimodal affective computing using data fusion and pattern classification techniques. Subsequently, we will develop learning, adaptive, and predictive models that allow for elements within the smart environment to dynamically respond to the affective states of users. Finally, we will study the impact of this paradigm on users and will model the performance of this human-in-the-loop smart environment over time. Our work will significantly contribute to the development of intelligent algorithms and systems such as wearables, IoT devices, and smart homes. Additionally, ubiquitous analysis and understanding of emotions and mental states of users, and developing responsive, adaptive, and predictive methods for interacting with users and influencing their affective states, will result in better products, user experience, and eventual quality of life, which will ultimately benefit Canadians.