User-Adapting System Design for Improved Energy Efficiency During the Use Phase of Products: Case Study of an Occupancy-Driven, Self-Learning Thermostat

Nonparametric user activity modelling and prediction

“…Finally, in real-time mode, the cross-correlations between the stored proles and the real-time detected ones are computed dynamically throughout the day and the most similar prole is selected. A non-parametric clustering method, Dirichlet process mixture, is developed in [58] and extended with a weekday variable in [12]. Prediction is then facilitated by selecting the most appropriate prole based on a prole's prior probability, pattern and weekday probability.…”

“…Introducing intelligence into devices, by modelling their usage, creating so-called intelligent products, is one avenue to achieve energy savings. Examples are found in a variety of application areas such as smart heating systems [2,12,17,24,33,36,38,39,41,43,53,64], which already found their way to the market (e.g. Nest, Heat Genius, Anna), appliance usage forecasting [3,4,8,20,29,51,60,67], dynamic power management [35,45,46,55], smart charging of electric vehicles [28,37] and fault detection [7,40,49,68].…”

Design for reduced resource consumption during the use phase of products

“…As part of the PERPETUAL project [169], De Bock et al (2016a) present a smart thermostat that establishes and exploits a user model to predict future occupancy and steer the heating accordingly. The resulting potential energy reduction of 259.2 kWh per year corresponds to 25.5% of the total energy consumption for a single---user office in Belgium [50]. Moreover, as user conduct changes over time, the profiling method was adapted in order to handle drifting behavior [51].…”

“…Therefore, some studies introduce a discomfort measure such as 'miss time' in smart heating systems. Miss time then represents the fraction of time the user experienced inconvenience as the model falsely predicted user absence and thus did not turn on, e.g., heating [50]. Finally, many of the reported potential savings are based on simulations.…”

“…As user behavior and preferences evolve over time, the systems must be able to detect and adapt to these changes. De Bock et al (2016a) presented an approach to deal with dynamic behavior for smart heating systems [50]. Again, it is essential that approaches are validated in real---life experiments.…”