Editorial to Special Issue on Smart Environments

摘要

Smart environments are sometimes regarded as ubiquitous systems comprising sensors, actuators, interaction components and computational elements connected as a network - but this picture is misleading just like picturing a smartphone merely as a mobile computing device would ignore its vital role within modern society for communication, organisation of business and daily life, and entertainment. The area of smart environments should thus be seen as installations supporting and enhancing the abilities of humans in their regular life as well as improving the environments themselves, for example in terms of energy efficiency. Smart environments are based on the technical substrate sketched above but they bear more challenges than the seamless composition of its components addressed in current research in technical disciplines. Setting technical challenges aside, developing intuitive interfaces to a system hidden in the environment and enabling the system to provide suitable assistance for a wide range of every-day tasks necessitates answers to manifold research questions. Indeed, at present smart environments are best pictured as an active field of research spanning several disciplines. Smart environments spans across a whole variety of core AI questions, but already a limited degree of 'smartness' can present a useful enhancement in some application contexts. This renders smart environments not only an interesting application area for contemporary AI research, but the challenges faced here are important stages towards long-term AI goals. Moreover, a careful design of interaction is necessary when developing smart environments to make their benefits intuitive and accessible to users. Designing effective modes of interaction has challenges in its own right, exceeding modalities and paradigms of classical approaches to human-machine interaction by far. For example, natural interaction with smart environments needs to be situated, that is, aware of and responsive to environmental and cognitive states. The ubiquitous nature of smart environments calls for implicit means of interaction. Moreover, systems capable of autonomous actions may shift initiative in interaction.