Outdoor Location Fingerprint Based on Optimized K-means Clustering

Abstract

In the indoor scene, location fingerprint has been developed to improve the locating speed by using clustering algorithm to process the fingerprint database. In the outdoor scene, location fingerprint is still at the stage of constructing a fingerprint database due to the vast region and its large data, so its locating speed needs to be faster. In the view of the slow speed, this paper proposes optimized K-means algorithm by ICFSFDP algorithm to process the outdoor fingerprint database, aiming to greatly increase the locating speed without sacrificing the accuracy of locating. In order to verify the performance of the algorithm, the standard propagation model is applied to calculate the sampling point’s reference signal received power (RSRP), and universal Kriging algorithm is used to interpolate the database, ensuring the authenticity of outdoor environment simulation. The result shows that the locating speed of outdoor fingerprint database can be greatly improved without decreasing accuracy, after being processed by I-CFSFDP optimized K-means.