Scattering Clustering Method for Terminal Fingerprint Positioning Based on MIMO Base Station
With the rapid development of wireless communication technology, location-based services are playing an increasingly important role in people’s lives. However, as the living environment becomes more and more complex, the existence of obstructions and various scatterers makes the accuracy of traditional positioning algorithms decrease, thus, fingerprint positioning has gradually become a research hotspot in the field of positioning. This paper researches the 5th Generation (5G) fingerprint location method based on machine learning. A massive multiple-in multiple-out (MIMO) channel is constructed on the MATLAB simulation platform, from which the fingerprint information is extracted to establish a fingerprint database. Considering the huge amount of data in the fingerprint database, and under the multipath effect, the channel characteristics are mainly affected by the scatterers near the point to be located. This paper proposes a scattering-based clustering method that combines the particularity of multipath propagation for clustering. Research shows that this method has excellent clustering effects, which can effectively improve algorithm efficiency and reduce data storage pressure on the base station side. (Foundation items: Social Development Projects of Jiangsu Science and Technology Department (No.BE2018704).)