Research On Indoor Wi-Fi Fingerprint Positioning Algorithm

With the rapid development of global navigation satellite system(GNSS)and wireless network technology,the demand for location-based services(LBS)has shown a substantial growth trend.Since people spend most of their time indoors,the demand for indoor location services has surged,making indoor positioning technology a focus of industrial applications and academic research hotspots.Due to the high popularity of WiFi,the Wi-Fi fingerprint positioning method is cheap and easy to implement,so the positioning method is widely used.However,Wi-Fi signals are easily affected by the environment,which affects the performance of Wi-Fi positioning.In response to the above problems,this article has launched the relevant research on indoor Wi-Fi fingerprint positioning technology.The specific research content includes the following points:(1)First of all,this thesis studies fingerprint positioning methods at home and abroad,analyzes the characteristics of Wi-Fi signal propagation in space,and summarizes the existing problems in fingerprint positioning.It provides a theoretical basis for the follow-up research of this article.(2)For large indoor environments,existing fingerprint positioning algorithms often combine clustering to reduce online positioning time and improve positioning accuracy,but the commonly used clustering algorithms are related to the number of clusters and initial cluster center.In addition,the wireless signal has a multipath effect,which leads to only relying on the signal domain distance measurement between fingerprints cannot accurately reflect the neighbor relationship between fingerprints.Therefore,an improved affine propagation clustering fingerprint positioning algorithm is proposed in this thesis,which uses the hybrid normalized distance as a similarity measure to perform affine propagation clustering on offline data.The estimated position of the test points are obtained by the inverse distance weighting method in the category which they belong.Experimental simulations show that the algorithm is suitable for a wide range of indoor environments.Compared with some existing algorithms,the positioning accuracy was improved and it didn’t need to specify the initial clustering center.(3)As the Wi-Fi signal changes with environmental changes,there are differences in fingerprints collected at different times at the same physical location.This leads to increased positioning errors,and the problem of re-collecting offline fingerprints is timeconsuming and labor-intensive.In a large-scale indoor environment,this problem is even more prominent.This thesis proposes a fingerprint location algorithm combined with transfer learning,which corrects the offline fingerprint database through the transfer learning algorithm.It can be concluded that the positioning performance is improved without the need to re-collect fingerprint data to update the fingerprint database.And the algorithm can reduce the impact of environmental changes on the positioning performance.