Indoor Positioning Based on Kernel Function and Analysis of Measurement Uncertainty

2nd International Symposium on Application of Materials Science and Energy Materials

材料科学;能源学

Wang, Yunling^1 ; Sang, Yu^1 ; Xu, Jie^1

Division of Electricity and Optics Measurement Technology, Shanghai Institute of Measurement and Testing Technology, China^1

关键词: Analysis of measurements;    Gaussian kernel functions;    Indoor positioning;    Matlab simulations;    Measurement uncertainty;    Non-linear correlations;    Received signal strength;    Reference points;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/490/6/062070/pdf
DOI  :  10.1088/1757-899X/490/6/062070

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

This paper proposes an indoor positioning algorithm based on kernel function, aiming at the problem that error of indoor positioning is large. Gaussian kernel function can capture the nonlinear correlation of received signal strength (RSS) between reference point fingerprint and the test point, and has a better matching effect than the KNN algorithm. The algorithm was verified by MATLAB simulation and real scene testing. At the end of the article, measurement uncertainty is analyzed to describe the reliability of the measurement results.