The intelligent thermostat
David V. Keyson, M. P. A. J. de Hoogh, Adinda Freudenthal, Arnold P. O. S. Vermeeren · 2000
Recent advances in approaches to mixed-initiative user interfaces for desktop applications provide a number of principles, which can be used to embed intelligence in consumer products. An integration of several approaches is proposed towards building an intelligent thermostat that can reduce energy consumption. A statistical model is used to spot trends in living patterns and make suggestions. An embedded task model, containing product usage patterns, can infer user intentions and serves as input to a dialog manager. Future work includes field-testing of the prototype and concepts in residential homes.