Research On Positioning Technology Based On Single Base Station With Multiple Antenna System

With the rapid development of wireless sensor communication technology,many location-based services(LBS)such as road navigation and target tracking had emerged.LBS is the antecedent demand in the smart city,smart home and intelligent transportation.In indoor positioning,multiple base stations are often set up for joint positioning,This method is called multi-base station positioning.Each base station which is in different position has a different perception of the environment,resulting in high fingerprint data feature differentiation and high positioning accuracy.However,in the aspect of single base station positioning,the research of single base station positioning is relatively lacking.Based on this,this dissertation conducts experimental research under the condition of single base station multi-antenna,and discusses the positioning performance of single base station multi-antenna.The main work is as follows:Firstly,the overall framework of single base station multi-antenna positioning model is designed,including the definition of module function and positioning process.MIMO base station equipment with 8 antennas was used to collect experimental data in a positioning area of 1500m~2.Secondly,based on the experience of multi-base station positioning,this dissertation analyzes the multi-base station positioning framework and algorithm scheme commonly used in the multi-base station positioning,and applies the algorithm scheme that is modified to the single-base station multi-antenna positioning system.Then,in order to solve the problem of hard boundary in the manual division of DNN region,a scheme based on DNN boundary sub-region detection+WKNN positioning algorithm was proposed,and the DNN model structure was modified to detect the boundary sub-region.The DNN model can determine the location of the undetermined site more accurately through two outputs,one outputs which region the undetermined site is in,and the other outputs whether the undetermined site is in the boundary sub-region,thus improving the positioning accuracy.Finally,the experiment is carried out under the single base station multi-antenna positioning system.Through different experiments,the results show that the DNN boundary sub-region detection+WKNN positioning algorithm scheme is better than the other three algorithms,and the average positioning error of about 5.2m is reached in the plane space of1500m~2 in the positioning area.