Research On Space Occupancy Detection For Smart Buildings Based On Multi-sensor Data Fusion Technology

The development of technologies such as Internet of Things,cloud computing has promoted the transformation of traditional buildings into smart buildings.However,the issue of enormous energy consumption in smart buildings also needs to be solved urgently.Nowadays,the key to achieving building energy efficiency lies in how to allocate building energy based on occupancy information.Therefore,in smart buildings,obtaining occupancy information of specific areas is of great significance for achieving energy-saving goals.Currently,mainstream occupancy perception solutions often use a single data as the information source.However,since the accuracy of the occupancy perception is highly dependent on the reliability of data,anomalous data collected can seriously affect the results of occupancy perception.Considering that a single source is no longer able to meet the high reliability and accuracy requirements of personnel perception solutions,this paper uses multi-sensor data fusion for occupancy perception and validates the effectiveness of this method through experiments.In addition,in the entire process of personnel occupancy perception,it is also very important to effectively evaluate the quality of data sources and fuse reliable and representative environmental factors.Therefore,the paper proposes two high-performance solutions for the above two aspects.The main work content carried out in this article is as follows:(1)In order to improve the WSN’s ability of identifying abnormal data,this paper proposes a plan to identify abnormal values from the dataset,which collected by the sensor network from two dimensions(horizontal and vertical detection).The experiment shows that this plan effectively improves the accuracy of anomaly detection,and provides standard correction objects(abnormal values)and reference data(normal values)for subsequent anomaly correction work.(2)To solve the problem that it is difficult to obtain reliable environmental factors,which can describe the entire area from a large amount of sensor data in a large area.This article first proposes a prediction model based on improved RBF,the model can correct and process the detected abnormal values,thus improving the reliability of environmental factors.Secondly,in order to obtain reliable and representative environmental factors,the corrected data are level-one fused through an improved normalized weighting averaging algorithm.(3)To address the problem of determining occupancy detection results through multiple heterogeneous data sources,this article introduces the D-S evidence theory into the evaluation of multiple occupancy information(level-two fusion)and fuses heterogeneous data sources to obtain high-precision occupancy information.Additionally,the article combines the Analytic Hierarchy Process(AHP)to solve the high-conflict problem of the D-S evidence theory,further improving the accuracy and robustness of occupancy detection results.