The Research Of WLAN Indoor Positioning Based On Hybrid Algorithm

With the rapid development of urbanization process in China, more and more new cities can be built and the rate of Chinese wireless network coverage is higher and higher so that the wireless services(LBS) which should base on the location of the indoor space would be increasingly widely applied to various shopping malls, nursing homes, residential as well as underground parking. As one of the research direction of LBS, which has been set up from an indoor WLAN launching point, it is an indoor wireless location positioning method as hardware device. More and more people pay attention on the WLAN position, it has also become an important development branch of the LBS. In this paper, depending on the existing indoor WLAN positioning technology, an indoor WLAN positioning method is proposed, which is effectiveness, robustness and high precision.In this paper, the WLAN signal in the indoor environment and the specific features of its channel propagation model are analyzed. We could explore and analyze several indoor positioning method in WLAN systems. Finally basing on the comparison of cost of performance comparative law, a based indoor signal Fingerprint point(RSS) positioning program is decided. This scheme can effectively avoid the practical difficulties in the use of TOA, TDOA( orientation programs often encounter, such as: the need to have all the target nodes and simultaneously anchor node high complexity and high cost required time synchronization issues brought), so that the positioning algorithm has a higher actual operability.Researchers in the research process based indoor WLAN positioning found the main reasons causing not high indoor WLAN positioning system location accuracy including aspects: WLAN signal in the communication process will always be affected on multipath effects(Multipath) and non-line interference(NLOS). Therefore, the fingerprint database which is set up by people cannot reflect the relationship between the true signal and the positional, but it was observed that the error value satisfy a Gaussian distribution characteristics in most cases. In this situation, this study could collect fingerprint signal and eliminate noise by median filtering. And then an indoor WLAN signal fingerprint database is established using the improved K-MEANS algorithm. Combining the two algorithms can make optimized indoor WLAN positioning method in a short period of time achieve target location and higher accuracy. Through further research, its actual positioning error values satisfy a Gaussian distribution in most cases, but in rare cases it may also fall on the index, Rayleigh and other distributions, but in practice after only 2-3 times faster positioning it can be ruled out. Therefore, an indoor WLAN positioning method ba sed on combining improved K-MEANS algorithm and median filtering is proposed, which has excellent robustness as well as the positioning accuracy is improved.