Study On Occupancy Using Wi-Fi Based Indoor Positioning System And Data Mining Methods

Detailed visualisation and data analysis of occupancy patterns including spatial distribution and temporal variations are of great importance to delivering energy efficient and productive buildings.The traditional occupancy positioning technologies often fail to detect the sedentary occupants inside buildings;by contrast,the Wi-Fi based positioning system can overcome this technical weakness.As an early explorative study,the experiment comprising 24-hour monitoring over 30 full days was conducted in the library of University of Reading,UK.Different from previous studies on occupancy and building energy consumption,specific occupancy information in terms of temporal and spatial characteristics can provide a depth understanding of occupants’ behaviour inside buildings,which could offer suggestions for building space optimism and building management.In addition,the indoor positions information related to occupants could offer implications for optimising space use,opening hours as well as staff deployment.Due to extremely high speed of gathering data by adopted equipment,a great volume of data was obtained.As a result,extensive work of data analysis including data pre-handling and choosing proper algorithms of data mining was required for effective pattern identification,to discuss the issues of building energy consumption and space utilization,through analysis the occupancy distribution in time and space.The characteristics of temporal distribution of occupancy is obtained after pre-handling;then,the association rule mining was used to extract the energy waste pattern,if this electricity waste pattern could be avoided,the consumed energy could be expected to be saved by 26.1%.With the assistance of clustering analysis,4 different occupancy patterns were identified.This work extracted the pattern of occupancy durations,and also the pattern of repeat visitors through the whole test period.In addition,the decision tree model is used to discuss the underlying effects of factors on occupancy duration.Future work is suggested to extend to more other types of rooms with diverse functions,other seasons and more different types of non-domestic buildings for a more comprehensive understanding of occupants’ behaviour and its underlying effects on building energy consumption.