| Global positioning system has been widely used in outdoor positioning applications,but it cannot be used in indoor environments where satellite signals are blocked.To solve this problem,researchers have developed various techniques for indoor positioning.Indoor positioning technology can be divided into two categories: active(requiring the person being positioned to hold the smart device)and passive(no need the person to be positioned holding the smart device).Compared with active positioning technology,passive positioning is non-invasive and requires no equipment.It can meet the needs of some emerging applications.For example,exhibition halls and shopping centers hope to support anti-theft services.In hospitals,medical personnel need to know the distribution location of patients who do not have portable equipment,but also in the home environment,indoor passive positioning can also be used for intrusion detection and so on.Because the WIIFI is densely deployed in many indoor environments,its application constraints are small,and it is possible to complete positioning without being perceived by the object,and it is feasible and advantageous to apply it to indoor passive positioning.So this article chooses the indoor passive positioning technology based on WIFI as the research content.Compared with RSSI(Received Signal Strength Indication),CSI(Channel State Information)can provide more detailed and finer channel information.This information may help to improve the accuracy of positioning technology,and CSI indicators are currently available on some commercial network cards.Acquired,not only limited to special equipment,making its application scenarios can be promoted.This paper compares and analyzes the effect of CSI and RSSI indicators on the positioning.After determining the feasibility,CSI is used as an indoor passive positioning indicator,and a multi-antenna system is designed to design a positioning technology that uses CSI space diversity to improve indoor passive positioning accuracy..Firstly,the application of CSI amplitude in indoor passive location was studied.In the data collection,a multi-antenna system with one transmit antenna and three receive antennas was used to collect CSI amplitude.By dividing the localized area into small units,the continuous positioning problem is transformed into a discrete classification problem,so that a multilayer feedforward neural network is selected to learn and train the CSI fingerprint data,and a classification model is obtained.When online forecasting,the prediction results of multiple samples are integrated into a data cluster by the voting mechanism,and the result with the most votes is the final positioning prediction result of the data cluster,so as to improve the accuracy of indoor passive positioning.The experiment also simulated the influence of several parameters including the number of antennas,data cluster size,and cell area on the indoor positioning performance,and objectively evaluated the effectiveness of the algorithm.The experimental results show that when the side length of the unit is 0.5m*0.5m,the data cluster size is 6,and the number of antennas is 3,the 97.6% positioning error is within 0.5m.Compared with similar algorithms,it is not only more accurate,but also the time and number of antennas requaried are less and there is better applicability.On this basis,this paper further analyzes the possibility of using phase as a fingerprint.Firstly,the noise of the original phase is removed by linear transformation,and then the processed phase combination amplitude is used as a new fingerprint feature.A better algorithm than the pure amplitude-based performance is obtained.The experimental results show that when the data cluster size is 1,and Compared with the simple amplitude-based algorithm,the probability that the positioning accuracy is within 0.5m is 90.%,which is an increase of 4%,while the algorithm using the original phase combined amplitude has a probability of only 83.3% when the positioning accuracy is 0.5m.The experiment verifies the feasibility of introducing the phase into the indoor passive positioning field. |
PDF Full Text Request