Occupancy detection and estimation in buildings paves the way to improve the utilization of lighting and HVAC systems,induce energy savings and enhance the well-being of the occupants.This paper presents a comparative...Occupancy detection and estimation in buildings paves the way to improve the utilization of lighting and HVAC systems,induce energy savings and enhance the well-being of the occupants.This paper presents a comparative study of state-of-art machine learning techniques that solve two different occupancy monitoring problems using environmental sensor data.One is the regression problem that estimates the actual count of occupants while the other is the classification problem which estimates the level of occupancy(empty,sparse,full).The results of the best performing machine learning techniques that solve both problems for the open dataset from the Uni-versity of Southern Denmark,Odense are presented to compare the accuracy of both approaches and the ease of implementation.The impact of𝐶𝑂2,temperature,and humidity features on the occupancy count/levels and de-tection accuracy(occupied versus unoccupied)are studied.Comprehensive analysis with different combinations of environmental features and other free features such as time-of-day along with different sampling techniques for training and testing are performed to understand how such models can be adapted for actual deployment.Our results indicate detection accuracy between 66%to 82%for different sampling schemes;with day-based sampling showing a better performance while random sampling generically showcasing lower accuracy(66.2%).The occupancy estimation(levels or counts)has accuracy in the range of 69%to 79%for random sampling and 71%to 80%for day-based sampling.Finally,results demonstrate that models based single environmental sensor data streams do not perform as well as the models with sensor fusion.展开更多
With the existence of several conventional and advanced building thermal energy demand forecast models to improve the energy efficiency of buildings,it is hard to find an appropriate,convenient,and efficient model.Eva...With the existence of several conventional and advanced building thermal energy demand forecast models to improve the energy efficiency of buildings,it is hard to find an appropriate,convenient,and efficient model.Evaluations based on statistical indexes(MAE,RMSE,MAPE,etc.)that characterize the accuracy of the forecasts do not help in the identification of the efficient building thermal energy demand forecast tool since they do not reflect the efforts entailed in implementation of the forecast model,i.e.,data collection to production/use phase.Hence,this work presents a Gini Index based Measurement of Alternatives and Ranking according to COmpromise Solution(GI-MARCOS),a hybrid Multi Attribute Decision Making(MADM)approach for the identification of the most efficient building energy demand forecast tool.GI-MARCOS employs(i)GI based objective weight method:assigns meaningful objective weights to the attributes in four phases(1:pre-processing,2:implementation,3:post-processing,and 4:use phase)thereby avoiding unnecessary biases in the expert’s opinion on weights and applicable to domains where there is a lack of domain expertise,and(ii)MARCOS:provides a robust and reliable ranking of alternatives in a dynamic environment.A case study with three alternatives evaluated over three to six attributes in four phases of implementation(pre-processing,implementation,post-processing and use)reveals that the use of GI-MARCOS improved the accuracy of alternatives MLR and BM by 6%and 13%,respectively.Moreover,additional validations state that(i)MLR performs best in Phase 1 and 2,while ANN performs best in Phase 3 and 4 with BM providing a mediocre performance in all four phases,(ii)sensitivity analysis:provides robust ranking with interchange of weights across phases and attributes,and(iii)rank correlation:ranks produce by GI-MARCOS has a high correlation with GRA(0.999),COPRAS(0.9786),and ARAS(0.9775).展开更多
基金supported by The Indian Institute of Technology-Bombay(Institute Postdoctoral Fellowship-AO/Admin-1/Rect/33/2019).
文摘Occupancy detection and estimation in buildings paves the way to improve the utilization of lighting and HVAC systems,induce energy savings and enhance the well-being of the occupants.This paper presents a comparative study of state-of-art machine learning techniques that solve two different occupancy monitoring problems using environmental sensor data.One is the regression problem that estimates the actual count of occupants while the other is the classification problem which estimates the level of occupancy(empty,sparse,full).The results of the best performing machine learning techniques that solve both problems for the open dataset from the Uni-versity of Southern Denmark,Odense are presented to compare the accuracy of both approaches and the ease of implementation.The impact of𝐶𝑂2,temperature,and humidity features on the occupancy count/levels and de-tection accuracy(occupied versus unoccupied)are studied.Comprehensive analysis with different combinations of environmental features and other free features such as time-of-day along with different sampling techniques for training and testing are performed to understand how such models can be adapted for actual deployment.Our results indicate detection accuracy between 66%to 82%for different sampling schemes;with day-based sampling showing a better performance while random sampling generically showcasing lower accuracy(66.2%).The occupancy estimation(levels or counts)has accuracy in the range of 69%to 79%for random sampling and 71%to 80%for day-based sampling.Finally,results demonstrate that models based single environmental sensor data streams do not perform as well as the models with sensor fusion.
基金supported by The Indian Institute of Technology-Bombay(Institute Postdoctoral Fellowship-AO/Admin-1/Rect/33/2019).
文摘With the existence of several conventional and advanced building thermal energy demand forecast models to improve the energy efficiency of buildings,it is hard to find an appropriate,convenient,and efficient model.Evaluations based on statistical indexes(MAE,RMSE,MAPE,etc.)that characterize the accuracy of the forecasts do not help in the identification of the efficient building thermal energy demand forecast tool since they do not reflect the efforts entailed in implementation of the forecast model,i.e.,data collection to production/use phase.Hence,this work presents a Gini Index based Measurement of Alternatives and Ranking according to COmpromise Solution(GI-MARCOS),a hybrid Multi Attribute Decision Making(MADM)approach for the identification of the most efficient building energy demand forecast tool.GI-MARCOS employs(i)GI based objective weight method:assigns meaningful objective weights to the attributes in four phases(1:pre-processing,2:implementation,3:post-processing,and 4:use phase)thereby avoiding unnecessary biases in the expert’s opinion on weights and applicable to domains where there is a lack of domain expertise,and(ii)MARCOS:provides a robust and reliable ranking of alternatives in a dynamic environment.A case study with three alternatives evaluated over three to six attributes in four phases of implementation(pre-processing,implementation,post-processing and use)reveals that the use of GI-MARCOS improved the accuracy of alternatives MLR and BM by 6%and 13%,respectively.Moreover,additional validations state that(i)MLR performs best in Phase 1 and 2,while ANN performs best in Phase 3 and 4 with BM providing a mediocre performance in all four phases,(ii)sensitivity analysis:provides robust ranking with interchange of weights across phases and attributes,and(iii)rank correlation:ranks produce by GI-MARCOS has a high correlation with GRA(0.999),COPRAS(0.9786),and ARAS(0.9775).
