Indoor location technology based on channel state information (CSI) has been widely used in various places

in the CSI phase fingerprint location scheme

the density of sampling points is closely related to the location effect.Due to the workload of data collection

previous scholars paid more attention to low-sampling density scenarios or adopted simulation methods to analyze sampling density research

which made it difficult to find the sampling density with the smallest positioning error.In this paper

the author first sets different CSI sampling densities

takes the CSI phase after data preprocessing as fingerprint characteristics

and then uses WKNN algorithm to analyze and explore the influence of different sampling densities on the positioning effect.The experiment results show that in the environment of 4 m×4 m

when a sampling point is taken every 0.4 m and the sampling density is 7.6 per square meter

the positioning accuracy is the highest

the average error is 0.34 m

and the sampling workload is taken into account.