Abstract | ||
---|---|---|
Indoor fingerprint location is widely used in passenger flow analysis, location based service, and security monitoring. RSS is usually used as signal fingerprint. However, in these applications the acquisition process relies on the uplink data transmission of the target device. These signals are often non-normally distributed and fluctuate greatly when the target moves. In order to cope with the above difficulties, this paper proposes a hidden markov model (HMM) framework based on targets maximum speed. It combines the speed parameters with path constraints to improve the calculation of state transition matrix. The kernel density estimation method based on the Wiener process is also used to solve the confusion matrix of HMM, which improves the accuracy of estimating the RSS distribution of the uplink signal. The performance of the algorithm are compared with the traditional HMM and NB methods in an actual indoor scenario. The influence of the target moving speed and the length of the observation sequence on the positioning accuracy is also analyzed. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1109/APCC.2018.8633478 | APCC |
Keywords | Field | DocType |
Hidden Markov models,Kernel,Estimation,Training,Probability density function,Monitoring,Uplink | Kernel (linear algebra),Confusion matrix,Data transmission,Computer science,Algorithm,Real-time computing,State-transition matrix,Hidden Markov model,RSS,Telecommunications link,Kernel density estimation | Conference |
ISBN | Citations | PageRank |
978-1-5386-6928-0 | 0 | 0.34 |
References | Authors | |
0 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Hao Chen | 1 | 156 | 61.18 |
Xiaofeng Tao | 2 | 1033 | 140.26 |
Yifan Zhang | 3 | 30 | 10.85 |
Wei Li | 4 | 422 | 56.67 |
zhang | 5 | 210 | 25.85 |