Fingerprinting Indoor Positioning Method Based on Kernel Ridge Regression with Feature Reduction

An important goal of indoor positioning systems is to improve accuracy as well reduce power consumption. In this paper, we propose an method based on the received signal strength (RSS) fingerprint. The proposed used a certain criterion select fixed access points (FPs) in offline phase instead online for location estimation. Principal component analysis (PCA) was applied features RSS measurements but retain most information possible establishing model. Then, kernel-based ridge regression obtain nonlinear relationship between principal components measures and position target. We thoroughly investigated performance realistic wireless local area network (WLAN) sensor (WSN) environments made comparisons with recently developed methods. experimental results indicated that less dependent density reference had higher than commonly methods, it adapts different application environments.