Research On Location Technology Of Fingerprint-based Crowdsourcing

With the rapid deployment of wireless local area networks(WLANs)and the widespread adoption of mobile smart terminals,there is a growing demand for Location Based Services(LBS).In indoor positioning technology,the location fingerprinting method based on Received Signal Strength(RSS)has the advantages of low cost,low power consumption,and easy implementation.It has become a research hotspot in indoor positioning.However,radio map construction in location fingerprinting will consume a lot of manpower,material resources and financial resources.The crowdsourcing model proposed in recent years has made the construction of radio maps simple and easy,saving a lot of labor costs.Therefore,crowd-based location fingerprinting technology has become a research hotspot in the field of indoor positioning today.In this paper,the access point(AP)selection,device heterogeneity and multi-floor scene location positioning based on crowd-source location fingerprinting are deeply studied.The main work is as follows:(1)Based on the basic theory research of indoor positioning,this paper analyzes the fingerprint positioning based on crowdsourcing,and builds a crowdsourcing-based positioning model,and analyzes the challenges faced by crowdsourcing-based fingerprint positioning.(2)AP selection scheme based on path loss model can effectively select APs with similar contribution to positioning,delete redundant APs,but do not consider AP stability,and Fisher criterion can achieve effective selection of stable APs.Therefore,in this paper,the Fisher criterion is introduced into the AP selection scheme based on the path loss model,and an improved path loss-based reliable AP selection(RAPC,Reliable AP Choice)scheme is proposed.In the early stage,the RSS fingerprint data was preprocessed,and the clustering algorithm was used to divide the sub-regions.At the later stage,AP selection was made using a reliable AP selection scheme.Finally,the K-Nearest Neighbor(KNN)algorithm was used to perform the positioning simulation experiment.Experiments show that the reliableAP selection scheme proposed in this paper has further improved the stability of the selected AP,making the constructed radio map more reliable,and thus improving the positioning accuracy.(3)In the heterogeneous solution based on similar sequences,the RSS library is used instead of the absolute RSS value to construct the fingerprint database,which improves the positioning accuracy under heterogeneous conditions,but does not consider the heterogeneity of the sensor.Therefore,this paper proposes an improved algorithm based on Similar Sequence and Sensor Fusion(SSSF).Particle filters are used to fuse similar sequences with sensors to track mobile users and mitigate the impact of heterogeneous devices.The simulation results show that the SSSF algorithm proposed in this paper can effectively solve the problem of heterogeneous equipment and further improve the positioning accuracy.(4)For multi-floor location problems,there is an existing algorithm that combines affine propagation clustering and logarithmic Gaussian likelihood functions to achieve the goal of transferring small databases from the network to the mobile terminal and improving real-time performance.The algorithm does not deal with non-zero data processing and generates incorrect clustering.In this paper,the algorithm is improved.Based on the logarithm Gaussian likelihood function,the Jaccard coefficient is used to correct the similarity,and the problem of the non-zero data processing before the improved algorithm is compensated.Comparing the improved algorithm proposed in this paper with the original algorithm and the traditional algorithm,the experiments show that the floor discriminant ratio,2D,3D positioning accuracy and the real-time positioning accuracy are all significantly improved.