Gaussian Process Regression Plus Method for Localization Reliability Improvement

• DOI: 10.3390/s16081193
• Publication Date: 2016-07-29T14:40:24Z
• Authors: Kehan Liu; Zhaopeng Meng; Chung-Ming Own
• Keywords: RSS; Data set

Location data are among the most widely used context in context-aware and ubiquitous computing applications. Many systems with distinct deployment costs positioning accuracies have been developed over past decade for indoor positioning. The useful method is focused on received signal strength provides a set of transmission access points. However, compiling manual measuring Received Signal Strength (RSS) fingerprint database involves high thus impractical an online prediction environment. system this study relied Gaussian process method, which nonparametric model that can be characterized completely by using mean function covariance matrix. In addition, Naive Bayes was to verify simplify computation precise predictions. authors conducted several experiments simulated real environments at Tianjin University. examined size, different kernels, accuracy. results showed proposed not only retain accuracy but also save time location