Research and improvement on indoor localization based on RSSI fingerprint database and K-nearest neighbor points

Abstract

Aiming at the shortcomings of K nearest neighbor algorithm, this paper put forward an indoor location algorithm based on K nearest neighbor collection of reference points. The new algorithm in this paper expand the single relationship between test points and reference points to net relationship between test points and reference points and between test points' close neighbor points and other reference points. The new algorithm uses the deeper information, and effectively reduces the influence of noise points. The new algorithm optimizes the formula of coordinate estimation through the occurrences of reference point. Experiments show that compared with K nearest neighbor localization algorithm, the new algorithm has improved on the positioning accuracy and stability.