Research On WiFi Indoor Localization Technology Based On Crowdsensing

With the sharp rise of mobile Internet and mobile terminal technology,people use the location-based services and feel the convenience.Considering that WiFi signals are already distributed throughout the indoor environment and commercial off-the-shelf mobile terminals can collect WiFi signals without adding new hardware devices,indoor localization based on WiFi is widely studied and used.The existing main algorithms can be divided into two categories:ranging based algorithms and non-ranging algorithms.The former relies on the distance between the unknown node and the reference nodes of known locations to obtain the location information;the latter obtains the location information by referring to the data in the database which is formed by the locations of known nodes and the fingerprints.Before localization,both need a lot of preparations.The former needs to obtain a map of indoor wireless access points,and the latter needs to collect the fingerprint database.The paper can obtain the relative positional relationship of the indoor nodes in real time without preparations;and 80%of the positioning accuracy is 6 meters with the offline data.Based on the concept of crowdsensing,this paper designs a signal strength quantization algorithm to quantify the similarity between received WiFi lists.In the simulation environment,the regression function between the value of quantization and the walking distance is calculated with the density of the wireless access point changed,and the squared value of the correlation coefficient~2 can reach 0.9316.In the experimental environment,the squared value of the correlation coefficient~2 is 0.779,which indicates strong correlation.The walking distances can be calculated based on the value of quantization and form a generalized distance matrix.By using the multidimensional scaling method,the generalized distance matrix in multidimensional space is reduced to the coordinate relationship in two-dimensional space.In this way,the nodes can be used as location fingerprints,and collecting the WiFi receiving lists at the nodes can form a partial indoor map,which greatly saves manpower and material resources.However,the accuracy of the signal strength quantization algorithm is degraded with the increase of distance.This paper designs a map stitching algorithm.It stitches the partial maps into a complete indoor map.Then the stretched complete indoor map is compared with the actual map,and 80%of the positioning accuracy is 6 meters,which is more accurate than the 7.5 meters obtained by the nearest neighbor algorithm of the same data amount.The complete indoor map of the paper is automatically generated without the wireless access point locations and actual indoor maps.