Research On Bluetooth Fingerprint Indoor Fingerprint Algorithm

Nowadays,with the rapid development of network technology,smart devices are becoming more and more popular,and location-based needs are becoming more and more important in people's lives.At present,outdoor positioning with GPS positioning technology as the core has been widely used in geographic data collection,vehicle monitoring and scheduling,mass consumption and other fields,but the complexity of indoor environment makes GPS and other commonly used positioning technologies prohibitive.This requires the selection of more suitable positioning technology in the indoor scene.At present,a variety of positioning techniques have been tried indoors,but the results are not ideal.With the advent of low-cost,lowpower Bluetooth,it has quickly become a research hotspot for indoor positioning.So everything is built around Bluetooth technology.The main research work and innovations of this paper are as follows:(1)Since the actual indoor environment is very different from the outdoor environment,a series of environmental factors such as people,objects,walls,etc.will cause signal reflection,refraction,and diffraction to occur,and thus to a certain extent,the establishment of the offline fingerprint location library is huge.influences.Therefore,this paper first collects the real indoor Bluetooth RSSI(Received Signal Strength Indicator)data,and simply studies the influence of distance,obstacles,human traffic and Bluetooth AP number on Bluetooth RSSI,so that there is a certain way to optimize the fingerprint database data.Understand and ensure the effectiveness of the generated training data and fingerprint maps,thus improving the positioning accuracy of the online positioning stage.(2)In the stage of offline fingerprint database establishment,this paper improves the original DBSCAN algorithm from two aspects.First,for the selection of parameters during abnormal point culling,the control variable method is used to continuously change the scan radius eps and the number of scans minPts to conduct a large number of experiments to select the parameter values more suitable for the test scene.Second,after determining a core point,the original algorithm re-scans all the points in the neighborhood of the core point,resulting in low execution efficiency,improving the operation mode,and querying the point farthest from the core point in the neighborhood.Save a lot of time at some unnecessary points.The experimental results show that the optimized DBSCAN algorithm improves the execution efficiency and achieves the effect while maintaining the accuracy compared with the traditional DBSCAN algorithm.(3)In the online positioning stage,this paper improves the original algorithm from three aspects.Firstly,how to quickly calculate the distance between each reference point and the point to be measured when the location fingerprint database matches,this paper adopts an adaptive method to determine each reference point.Secondly,the selection of the K value of the reference point is studied through a large number of experiments to finally determine a more suitable K value.Then,the contribution to the K value is normalized by Euclidean distance,which ensures the rationality of the contribution of each reference point.Finally,the improved normalized adaptive WKNN algorithm is verified by the actual environment.The experimental results show that the improved normalized adaptive WKNN algorithm has improved the positioning accuracy and execution efficiency.